Upload 14 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+data/emotions.mp4 filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,66 @@
+import streamlit as st
+import time
+import cv2
+from utils import emotion_detection_brainai as edb
+emotion_model = edb.EmotionModel()
+
+st.set_page_config(
+    page_title = "Emotion Detection", 
+    page_icon = "	:full_moon_with_face:",
+    layout = "wide")
+
+st.title(":rainbow[감정 인식] 	:full_moon_with_face:")
+st.sidebar.header("메뉴")
+source_radio = st.sidebar.radio("선택하세요", ["IMAGE", "VIDEO", "WEBCAM"])
+
+if source_radio == "IMAGE":
+    st.sidebar.header("이미지 파일 업로드")
+    img = st.sidebar.file_uploader("이미지 파일을 선택하세요.", type=("jpg", "png"))
+    st.write(":green[왼쪽 메뉴 'Browse files' 버튼을 클릭하여 이미지 파일을 선택하면 AI 추론이 시작됩니다.]" )
+    if img is not None:
+        result_img = emotion_model.process(img)
+        st.image(result_img)
+    
+    else:
+        image_name = [ "anger", "happy", "neutral", "sad",  "surprise",]
+        cols = st.columns(len(image_name)) 
+        for i, name in enumerate(image_name):
+            path = "data/" + name + ".jpg"
+            with cols[i]:
+                st.image(path, caption=name)
+    
+elif source_radio == "VIDEO":
+    
+    st.sidebar.header("비디오 파일 업로드")
+    input_video = st.sidebar.file_uploader("비디오 파일을 선택하세요..", type=("mp4"))
+    st.write(":green[왼쪽 메뉴 'Browse files' 버튼을 클릭하여 비디오 파일을 선택하면 AI 추론이 시작됩니다.]" )
+    if input_video is not None:
+        temp_file = edb.play_video(input_video)
+        st.video(temp_file)
+        
+    else:
+        st.video("data/emotions.mp4")
+
+elif source_radio == "WEBCAM":
+    video_capture = cv2.VideoCapture(0)  # 0 usually refers to the default webcam
+    placeholder = st.empty()
+    stop_streaming = st.button("Stop Streaming")  # No key needed here since it's only one button
+
+    while True:
+        ret, frame = video_capture.read()  # Read a frame from the webcam
+        
+        if not ret:
+            break  # Break the loop if there's an issue with the webcam
+         
+        result_img = emotion_model.process(frame)
+        result_img_rgb = cv2.cvtColor(result_img, cv2.COLOR_BGR2RGB)
+        placeholder.image(result_img_rgb)
+        time.sleep(0.03)  # Approximately 30 frames per second. Too low values can overload the browser
+        if stop_streaming:  # If the button is clicked, stop_streaming will be True
+            break
+
+    video_capture.release() # Release the webcam resources when done
+    st.write("Streaming stopped.")
+
+    
+ 

data/emotions.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4e84fe138ea681bcb5410fe62e7a7054d956b251173e90a02e34df0cba36d1e
+size 7624684

models/emotions-recognition-retail-0003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcb9b1a910fa3cd18a638bb1dbb0597c4ef7a080d1b83008c8e8c2c3c42b99dd
+size 9930028

models/emotions-recognition-retail-0003.xml

@@ -0,0 +1,1496 @@
+<?xml version="1.0" ?>
+<net name="0003_EmoNet_ResNet10" version="10">
+	<layers>
+		<layer id="0" name="data" type="Parameter" version="opset1">
+			<data element_type="f32" shape="1, 3, 64, 64"/>
+			<output>
+				<port id="0" names="data" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="1" name="data_mul_2237" type="Const" version="opset1">
+			<data element_type="f32" offset="0" shape="1, 3, 1, 1" size="12"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="2" name="BatchNorm1/mean/Fused_Mul_" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="3" name="data_add_2239" type="Const" version="opset1">
+			<data element_type="f32" offset="12" shape="1, 3, 1, 1" size="12"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="4" name="BatchNorm1/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="data" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="5" name="BatchNorm2/mean/Fused_Mul__copy" type="Const" version="opset1">
+			<data element_type="f32" offset="24" shape="64, 3, 7, 7" size="37632"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>7</dim>
+					<dim>7</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="6" name="Convolution1/WithoutBiases" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="3, 3" pads_end="3, 3" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>3</dim>
+					<dim>64</dim>
+					<dim>64</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>7</dim>
+					<dim>7</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="7" name="data_add_22422247" type="Const" version="opset1">
+			<data element_type="f32" offset="37656" shape="1, 64, 1, 1" size="256"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="8" name="Convolution1/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="9" name="ReLU1" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Convolution1,Dropout_ReLU1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="10" name="Pooling1" type="MaxPool" version="opset1">
+			<data auto_pad="explicit" kernel="3, 3" pads_begin="0, 0" pads_end="0, 0" rounding_type="ceil" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>32</dim>
+					<dim>32</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Pooling1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="11" name="BatchNorm3/mean/Fused_Mul__copy" type="Const" version="opset1">
+			<data element_type="f32" offset="37912" shape="64, 64, 3, 3" size="147456"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="12" name="Convolution2" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>64</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="13" name="data_add_22502255" type="Const" version="opset1">
+			<data element_type="f32" offset="185368" shape="1, 64, 1, 1" size="256"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="14" name="BatchNorm3/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution2" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="15" name="ReLU2" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Convolution2" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="16" name="125" type="Const" version="opset1">
+			<data element_type="f32" offset="185624" shape="64, 64, 3, 3" size="147456"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>64</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="17" name="Convolution3" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>64</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution3" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="18" name="Eltwise1" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="19" name="data_mul_22572261" type="Const" version="opset1">
+			<data element_type="f32" offset="333080" shape="1, 64, 1, 1" size="256"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="20" name="BatchNorm4/mean/Fused_Mul_" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="21" name="data_add_22582263" type="Const" version="opset1">
+			<data element_type="f32" offset="333336" shape="1, 64, 1, 1" size="256"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="22" name="BatchNorm4/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="23" name="ReLU3" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Dropout_ReLU3,Eltwise1" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="24" name="BatchNorm5/mean/Fused_Mul__copy" type="Const" version="opset1">
+			<data element_type="f32" offset="333592" shape="128, 64, 3, 3" size="294912"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>128</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="25" name="Convolution4" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>128</dim>
+					<dim>64</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="26" name="data_add_22662271" type="Const" version="opset1">
+			<data element_type="f32" offset="628504" shape="1, 128, 1, 1" size="512"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="27" name="BatchNorm5/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution4" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="28" name="ReLU4" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Convolution4" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="29" name="114" type="Const" version="opset1">
+			<data element_type="f32" offset="629016" shape="128, 128, 3, 3" size="589824"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>128</dim>
+					<dim>128</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="30" name="Convolution5" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>128</dim>
+					<dim>128</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution5" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="31" name="65" type="Const" version="opset1">
+			<data element_type="f32" offset="1218840" shape="128, 64, 1, 1" size="32768"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>128</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="32" name="Convolution6" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>64</dim>
+					<dim>16</dim>
+					<dim>16</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>128</dim>
+					<dim>64</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution6" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="33" name="Eltwise2" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise2" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="34" name="data_mul_22732277" type="Const" version="opset1">
+			<data element_type="f32" offset="1251608" shape="1, 128, 1, 1" size="512"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="35" name="BatchNorm6/mean/Fused_Mul_" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="36" name="data_add_22742279" type="Const" version="opset1">
+			<data element_type="f32" offset="1252120" shape="1, 128, 1, 1" size="512"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="37" name="BatchNorm6/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise2" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="38" name="ReLU5" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Dropout_ReLU5,Eltwise2" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="39" name="BatchNorm7/mean/Fused_Mul__copy" type="Const" version="opset1">
+			<data element_type="f32" offset="1252632" shape="256, 128, 3, 3" size="1179648"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>128</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="40" name="Convolution7" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>128</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="41" name="data_add_22822287" type="Const" version="opset1">
+			<data element_type="f32" offset="2432280" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="42" name="BatchNorm7/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution7" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="43" name="ReLU6" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Convolution7" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="44" name="91" type="Const" version="opset1">
+			<data element_type="f32" offset="2433304" shape="256, 256, 3, 3" size="2359296"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="45" name="Convolution8" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution8" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="46" name="89" type="Const" version="opset1">
+			<data element_type="f32" offset="4792600" shape="256, 128, 1, 1" size="131072"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="47" name="Convolution9" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>128</dim>
+					<dim>8</dim>
+					<dim>8</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>128</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution9" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="48" name="Eltwise3" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise3" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="49" name="data_mul_22892293" type="Const" version="opset1">
+			<data element_type="f32" offset="4923672" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="50" name="BatchNorm8/mean/Fused_Mul_" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="51" name="data_add_22902295" type="Const" version="opset1">
+			<data element_type="f32" offset="4924696" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="52" name="BatchNorm8/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise3" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="53" name="ReLU7" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Dropout_ReLU7,Eltwise3" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="54" name="BatchNorm9/mean/Fused_Mul__copy" type="Const" version="opset1">
+			<data element_type="f32" offset="4925720" shape="256, 256, 3, 3" size="2359296"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="55" name="Convolution10" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="56" name="data_add_22982303" type="Const" version="opset1">
+			<data element_type="f32" offset="7285016" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="57" name="BatchNorm9/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution10" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="58" name="ReLU8" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Convolution10" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="59" name="66" type="Const" version="opset1">
+			<data element_type="f32" offset="7286040" shape="256, 256, 3, 3" size="2359296"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="60" name="Convolution11" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>3</dim>
+					<dim>3</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution11" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="61" name="97" type="Const" version="opset1">
+			<data element_type="f32" offset="9645336" shape="256, 256, 1, 1" size="262144"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="62" name="Convolution12" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" strides="2, 2"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>4</dim>
+					<dim>4</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>256</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Convolution12" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="63" name="Eltwise4" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise4" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="64" name="data_mul_23052309" type="Const" version="opset1">
+			<data element_type="f32" offset="9907480" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="65" name="BatchNorm10/mean/Fused_Mul_" type="Multiply" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="66" name="data_add_23062311" type="Const" version="opset1">
+			<data element_type="f32" offset="9908504" shape="1, 256, 1, 1" size="1024"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="67" name="BatchNorm10/variance/Fused_Add_" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="Eltwise4" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="68" name="ReLU9" type="ReLU" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="Dropout_ReLU9,Eltwise4" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="69" name="94" type="Const" version="opset1">
+			<data element_type="f32" offset="9909528" shape="5, 256, 2, 2" size="20480"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>5</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="70" name="emotion5_conv1/WithoutBiases" type="Convolution" version="opset1">
+			<data auto_pad="explicit" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" strides="1, 1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>5</dim>
+					<dim>256</dim>
+					<dim>2</dim>
+					<dim>2</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="71" name="emotion5_conv1/Dims1316" type="Const" version="opset1">
+			<data element_type="f32" offset="9930008" shape="1, 5, 1, 1" size="20"/>
+			<output>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="72" name="emotion5_conv1" type="Add" version="opset1">
+			<data auto_broadcast="numpy"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+				<port id="1" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="2" names="emotion5_conv" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="73" name="prob_emotion" type="SoftMax" version="opset1">
+			<data axis="1"/>
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+			<output>
+				<port id="1" names="prob_emotion" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</output>
+		</layer>
+		<layer id="74" name="prob_emotion/sink_port_0" type="Result" version="opset1">
+			<input>
+				<port id="0" precision="FP32">
+					<dim>1</dim>
+					<dim>5</dim>
+					<dim>1</dim>
+					<dim>1</dim>
+				</port>
+			</input>
+		</layer>
+	</layers>
+	<edges>
+		<edge from-layer="0" from-port="0" to-layer="2" to-port="0"/>
+		<edge from-layer="1" from-port="0" to-layer="2" to-port="1"/>
+		<edge from-layer="2" from-port="2" to-layer="4" to-port="0"/>
+		<edge from-layer="3" from-port="0" to-layer="4" to-port="1"/>
+		<edge from-layer="4" from-port="2" to-layer="6" to-port="0"/>
+		<edge from-layer="5" from-port="0" to-layer="6" to-port="1"/>
+		<edge from-layer="6" from-port="2" to-layer="8" to-port="0"/>
+		<edge from-layer="7" from-port="0" to-layer="8" to-port="1"/>
+		<edge from-layer="8" from-port="2" to-layer="9" to-port="0"/>
+		<edge from-layer="9" from-port="1" to-layer="10" to-port="0"/>
+		<edge from-layer="10" from-port="1" to-layer="12" to-port="0"/>
+		<edge from-layer="10" from-port="1" to-layer="18" to-port="1"/>
+		<edge from-layer="11" from-port="0" to-layer="12" to-port="1"/>
+		<edge from-layer="12" from-port="2" to-layer="14" to-port="0"/>
+		<edge from-layer="13" from-port="0" to-layer="14" to-port="1"/>
+		<edge from-layer="14" from-port="2" to-layer="15" to-port="0"/>
+		<edge from-layer="15" from-port="1" to-layer="17" to-port="0"/>
+		<edge from-layer="16" from-port="0" to-layer="17" to-port="1"/>
+		<edge from-layer="17" from-port="2" to-layer="18" to-port="0"/>
+		<edge from-layer="18" from-port="2" to-layer="20" to-port="0"/>
+		<edge from-layer="19" from-port="0" to-layer="20" to-port="1"/>
+		<edge from-layer="20" from-port="2" to-layer="22" to-port="0"/>
+		<edge from-layer="21" from-port="0" to-layer="22" to-port="1"/>
+		<edge from-layer="22" from-port="2" to-layer="23" to-port="0"/>
+		<edge from-layer="23" from-port="1" to-layer="32" to-port="0"/>
+		<edge from-layer="23" from-port="1" to-layer="25" to-port="0"/>
+		<edge from-layer="24" from-port="0" to-layer="25" to-port="1"/>
+		<edge from-layer="25" from-port="2" to-layer="27" to-port="0"/>
+		<edge from-layer="26" from-port="0" to-layer="27" to-port="1"/>
+		<edge from-layer="27" from-port="2" to-layer="28" to-port="0"/>
+		<edge from-layer="28" from-port="1" to-layer="30" to-port="0"/>
+		<edge from-layer="29" from-port="0" to-layer="30" to-port="1"/>
+		<edge from-layer="30" from-port="2" to-layer="33" to-port="0"/>
+		<edge from-layer="31" from-port="0" to-layer="32" to-port="1"/>
+		<edge from-layer="32" from-port="2" to-layer="33" to-port="1"/>
+		<edge from-layer="33" from-port="2" to-layer="35" to-port="0"/>
+		<edge from-layer="34" from-port="0" to-layer="35" to-port="1"/>
+		<edge from-layer="35" from-port="2" to-layer="37" to-port="0"/>
+		<edge from-layer="36" from-port="0" to-layer="37" to-port="1"/>
+		<edge from-layer="37" from-port="2" to-layer="38" to-port="0"/>
+		<edge from-layer="38" from-port="1" to-layer="47" to-port="0"/>
+		<edge from-layer="38" from-port="1" to-layer="40" to-port="0"/>
+		<edge from-layer="39" from-port="0" to-layer="40" to-port="1"/>
+		<edge from-layer="40" from-port="2" to-layer="42" to-port="0"/>
+		<edge from-layer="41" from-port="0" to-layer="42" to-port="1"/>
+		<edge from-layer="42" from-port="2" to-layer="43" to-port="0"/>
+		<edge from-layer="43" from-port="1" to-layer="45" to-port="0"/>
+		<edge from-layer="44" from-port="0" to-layer="45" to-port="1"/>
+		<edge from-layer="45" from-port="2" to-layer="48" to-port="0"/>
+		<edge from-layer="46" from-port="0" to-layer="47" to-port="1"/>
+		<edge from-layer="47" from-port="2" to-layer="48" to-port="1"/>
+		<edge from-layer="48" from-port="2" to-layer="50" to-port="0"/>
+		<edge from-layer="49" from-port="0" to-layer="50" to-port="1"/>
+		<edge from-layer="50" from-port="2" to-layer="52" to-port="0"/>
+		<edge from-layer="51" from-port="0" to-layer="52" to-port="1"/>
+		<edge from-layer="52" from-port="2" to-layer="53" to-port="0"/>
+		<edge from-layer="53" from-port="1" to-layer="55" to-port="0"/>
+		<edge from-layer="53" from-port="1" to-layer="62" to-port="0"/>
+		<edge from-layer="54" from-port="0" to-layer="55" to-port="1"/>
+		<edge from-layer="55" from-port="2" to-layer="57" to-port="0"/>
+		<edge from-layer="56" from-port="0" to-layer="57" to-port="1"/>
+		<edge from-layer="57" from-port="2" to-layer="58" to-port="0"/>
+		<edge from-layer="58" from-port="1" to-layer="60" to-port="0"/>
+		<edge from-layer="59" from-port="0" to-layer="60" to-port="1"/>
+		<edge from-layer="60" from-port="2" to-layer="63" to-port="0"/>
+		<edge from-layer="61" from-port="0" to-layer="62" to-port="1"/>
+		<edge from-layer="62" from-port="2" to-layer="63" to-port="1"/>
+		<edge from-layer="63" from-port="2" to-layer="65" to-port="0"/>
+		<edge from-layer="64" from-port="0" to-layer="65" to-port="1"/>
+		<edge from-layer="65" from-port="2" to-layer="67" to-port="0"/>
+		<edge from-layer="66" from-port="0" to-layer="67" to-port="1"/>
+		<edge from-layer="67" from-port="2" to-layer="68" to-port="0"/>
+		<edge from-layer="68" from-port="1" to-layer="70" to-port="0"/>
+		<edge from-layer="69" from-port="0" to-layer="70" to-port="1"/>
+		<edge from-layer="70" from-port="2" to-layer="72" to-port="0"/>
+		<edge from-layer="71" from-port="0" to-layer="72" to-port="1"/>
+		<edge from-layer="72" from-port="2" to-layer="73" to-port="0"/>
+		<edge from-layer="73" from-port="1" to-layer="74" to-port="0"/>
+	</edges>
+	<meta_data>
+		<MO_version value="2021.4.0-3827-c5b65f2cb1d-releases/2021/4"/>
+		<cli_parameters>
+			<caffe_parser_path value="DIR"/>
+			<data_type value="FP32"/>
+			<disable_nhwc_to_nchw value="False"/>
+			<disable_omitting_optional value="False"/>
+			<disable_resnet_optimization value="False"/>
+			<disable_weights_compression value="False"/>
+			<enable_concat_optimization value="False"/>
+			<enable_flattening_nested_params value="False"/>
+			<enable_ssd_gluoncv value="False"/>
+			<extensions value="DIR"/>
+			<framework value="caffe"/>
+			<freeze_placeholder_with_value value="{}"/>
+			<generate_deprecated_IR_V7 value="False"/>
+			<input value="data"/>
+			<input_model value="DIR/0003_EmoNet_ResNet10.caffemodel"/>
+			<input_model_is_text value="False"/>
+			<input_proto value="DIR/0003_EmoNet_ResNet10.prototxt"/>
+			<input_shape value="[1,3,64,64]"/>
+			<k value="DIR/CustomLayersMapping.xml"/>
+			<keep_shape_ops value="True"/>
+			<legacy_ir_generation value="False"/>
+			<legacy_mxnet_model value="False"/>
+			<log_level value="ERROR"/>
+			<mean_scale_values value="{'data': {'mean': None, 'scale': array([1.])}}"/>
+			<mean_values value="()"/>
+			<model_name value="emotions-recognition-retail-0003"/>
+			<output value="['prob_emotion']"/>
+			<output_dir value="DIR"/>
+			<placeholder_data_types value="{}"/>
+			<placeholder_shapes value="{'data': array([ 1,  3, 64, 64])}"/>
+			<progress value="False"/>
+			<remove_memory value="False"/>
+			<remove_output_softmax value="False"/>
+			<reverse_input_channels value="False"/>
+			<save_params_from_nd value="False"/>
+			<scale_values value="data[1.0]"/>
+			<silent value="False"/>
+			<static_shape value="False"/>
+			<stream_output value="False"/>
+			<transform value=""/>
+			<unset unset_cli_parameters="batch, counts, disable_fusing, disable_gfusing, finegrain_fusing, input_checkpoint, input_meta_graph, input_symbol, mean_file, mean_file_offsets, move_to_preprocess, nd_prefix_name, pretrained_model_name, saved_model_dir, saved_model_tags, scale, tensorboard_logdir, tensorflow_custom_layer_libraries, tensorflow_custom_operations_config_update, tensorflow_object_detection_api_pipeline_config, tensorflow_use_custom_operations_config, transformations_config"/>
+		</cli_parameters>
+	</meta_data>
+</net>

models/face-detection-adas-0001.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54883903421c8ee5289334ca4da779976a8be66daf7d57e45d005d0bc2b8c637
+size 4212072

models/person-detection-retail-0013.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f09cb7061328942f9d5e9fc81631a4234be66a26daa50cd672d4077ee82ad44
+size 2891364

utils/emotion_detection_brainai.py

@@ -0,0 +1,136 @@
+import openvino as ov
+import cv2
+import numpy as np
+import PIL
+import io
+import tempfile
+import streamlit as st
+import moviepy.editor as mpy
+
+
+class EmotionModel:
+    def __init__(self):
+        self.face_compiled_model, self.face_input_layer, self.face_output_layer = self.load_model('face-detection-adas-0001')
+        self.emotion_compiled_model, self.emotion_input_layer, self.emotion_output_layer = self.load_model('emotions-recognition-retail-0003')
+        
+
+    def load_model(self, model_name):
+        model_path = "models/" + model_name + ".xml"
+        core = ov.Core()
+        model = core.read_model(model=model_path)
+        compiled_model = core.compile_model(model=model, device_name="CPU")
+        input_layer = compiled_model.input(0)
+        output_layer = compiled_model.output(0)
+        return compiled_model, input_layer, output_layer
+
+    def preprocess(self, img, input_layer):
+ 
+        input_h, input_w = input_layer.shape[2], input_layer.shape[3]
+        input_img = cv2.resize(img, (input_w,input_h))
+        input_img = input_img.transpose(2, 0, 1)
+        input_img = np.expand_dims(input_img, 0)
+        
+        return input_img
+
+
+    def post_process_face(self, result_face, img, conf=0.5):
+        boxes = []
+        h,w,_ = img.shape
+        predictions = result_face[0][0]       # 하위 집합 데이터 프레임
+        confidence = predictions[:,2]    # conf 값 가져오기 [img_id, label, conf, x_min, y_min, x_max, y_max]
+      
+        top_predictions = predictions[(confidence>conf)]         # 임계값보다 큰 conf 값을 가진 예측만 선택
+        for detection in top_predictions:
+            box = (detection[3:7]* np.array([w, h, w, h])).astype("int") # 상자 위치 결정
+            box = [0 if i < 0 else i for i in box]
+            (xmin, ymin, xmax, ymax) = box   # xmin, ymin, xmax, ymax에 상자 위치 값 지정
+            boxes.append(box)
+            cv2.rectangle(img, (xmin, ymin), (xmax, ymax), (0, 0, 255), 2)       # 사각형 만들기
+        
+        return boxes
+
+    def post_process_emotion(self, result_emotion, img, face_position):
+        
+        emotions = {
+            0:"neutral",
+            1:"happy",
+            2:"sad",
+            3:"surprise",
+            4:"anger"
+        }
+                    
+        predictions = result_emotion[0,:,0,0]   
+        topresult_index = np.argmax(predictions)
+        emotion = emotions[topresult_index]
+
+        font_size = img.shape[0]/1000
+        font_thickness = int(img.shape[0]/500)
+        text_offset = int(img.shape[0]/30)
+       
+        cv2.putText(img, emotion,
+                    (face_position[0],face_position[1]+text_offset),
+                    cv2.FONT_HERSHEY_SIMPLEX, font_size, 
+                    (255, 255,255), font_thickness)
+        
+        return emotion
+
+    def process(self, img):
+
+        if isinstance(img, np.ndarray):
+            uploaded_img_cv = img
+        else:
+            uploaded_img = PIL.Image.open(img)
+            uploaded_img_cv =  np.array(uploaded_img) 
+
+        input_img = self.preprocess(uploaded_img_cv, self.face_input_layer)
+        result_face = self.face_compiled_model([input_img])[self.face_output_layer]
+        boxes = self.post_process_face(result_face, uploaded_img_cv, conf=0.5)
+
+        if boxes is not None:
+
+            for box in boxes:
+                xmin, ymin, xmax, ymax = box
+                emotion_input = uploaded_img_cv[ymin:ymax,xmin:xmax]
+                input_img = self.preprocess(emotion_input, self.emotion_input_layer)       
+                result_emotion = self.emotion_compiled_model([input_img])[self.emotion_output_layer]
+                self.post_process_emotion(result_emotion, uploaded_img_cv, box)
+
+        return uploaded_img_cv
+
+emotion_model = EmotionModel()
+def play_video(input_video):
+    g = io.BytesIO(input_video.read())
+    temporary_location = "upload.mp4" 
+    with open(temporary_location, "wb") as out: 
+        out.write(g.read())
+    out.close() 
+    
+    camera = cv2.VideoCapture(temporary_location)
+    fps = camera.get(cv2.CAP_PROP_FPS)
+    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix='.mp4')
+    video_row=[]
+    total_frames = int(camera.get(cv2.CAP_PROP_FRAME_COUNT))
+    progress_bar = st.progress(0)
+    frame_count = 0
+
+    st_frame = st.empty()
+    while(camera.isOpened()):
+        ret, frame = camera.read()
+
+        if ret:
+            emotion_img = emotion_model.process(frame)
+            st_frame.image(emotion_img, channels = "BGR")
+            video_row.append(cv2.cvtColor(emotion_img,cv2.COLOR_BGR2RGB))
+            frame_count +=1
+            progress_bar.progress(frame_count/total_frames, text = None)
+
+        else:
+            camera.release()
+            st_frame.empty()
+            progress_bar.empty()
+            break
+    clip = mpy.ImageSequenceClip(video_row,fps=fps)
+    clip.write_videofile(temp_file.name)
+
+    return temp_file.name
+