First draft

README.md

@@ -1,13 +1,8 @@
----
-title: ASL Classifier
-emoji: 📊
-colorFrom: gray
-colorTo: blue
-sdk: gradio
-sdk_version: 5.49.1
-app_file: app.py
-pinned: false
-short_description: ASL Classifier App
----
+# 🖐️ ASL Letter Classifier (YOLOv11 + MediaPipe)
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This app uses:
+- **YOLOv11 Classification** (fine-tuned on ASL alphabet images)
+- **MediaPipe Hand Landmarks** for hand annotation
+- **Gradio Web UI** for uploading and testing images
+
+Upload a photo of a hand sign to see the detected letter and confidence.

app.py

@@ -0,0 +1,89 @@
+from ultralytics import YOLO
+import gradio as gr
+import cv2
+import os
+import numpy as np
+from mediapipe import Image
+from mediapipe.tasks import python
+from mediapipe.tasks.python import vision
+
+# ---------------------
+# Load YOLO model
+# ---------------------
+MODEL_PATH = "yolov11n_finetuned_ASL.pt"
+HAND_MODEL_PATH = "hand_landmarker.task"
+OUTPUT_DIR = "annotated_uploads"
+os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+model = YOLO(MODEL_PATH)
+
+# ---------------------
+# Load MediaPipe hand landmark detector
+# ---------------------
+base_options = python.BaseOptions(model_asset_path=HAND_MODEL_PATH)
+options = vision.HandLandmarkerOptions(base_options=base_options, num_hands=1)
+detector = vision.HandLandmarker.create_from_options(options)
+
+# ---------------------
+# Helper: Annotate image with landmarks
+# ---------------------
+def annotate_with_mediapipe(image_path):
+    img = cv2.imread(image_path)
+    if img is None:
+        return image_path  # fallback
+
+    mp_image = Image.create_from_file(image_path)
+    detection_result = detector.detect(mp_image)
+
+    if detection_result.hand_landmarks:
+        for hand_landmarks in detection_result.hand_landmarks:
+            for landmark in hand_landmarks:
+                h, w, _ = img.shape
+                x, y = int(landmark.x * w), int(landmark.y * h)
+                cv2.circle(img, (x, y), 3, (0, 255, 0), -1)  # green points
+
+    annotated_path = os.path.join(OUTPUT_DIR, os.path.basename(image_path))
+    cv2.imwrite(annotated_path, img)
+    return annotated_path
+
+# ---------------------
+# Prediction function
+# ---------------------
+def predict(image):
+    # Save uploaded image temporarily
+    temp_path = "temp_upload.jpg"
+    image.save(temp_path)
+
+    # Step 1: Annotate
+    annotated_path = annotate_with_mediapipe(temp_path)
+
+    # Step 2: Run YOLO prediction
+    results = model.predict(annotated_path, imgsz=300, verbose=False)[0]
+    probs = results.probs
+    top_idx = probs.top1
+    top_label = results.names[top_idx]
+    confidence = probs.data[top_idx].item()
+
+    # Step 3: Load annotated image for display
+    annotated_img = cv2.imread(annotated_path)
+    annotated_img = cv2.cvtColor(annotated_img, cv2.COLOR_BGR2RGB)
+
+    return annotated_img, {top_label: confidence}
+
+# ---------------------
+# Build Gradio Interface
+# ---------------------
+demo = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil", label="Upload a Hand Image"),
+    outputs=[
+        gr.Image(label="Annotated Image"),
+        gr.Label(label="Predicted Letter (Confidence)")
+    ],
+    title="🖐️ ASL Letter Classifier",
+    description="Upload an image of a hand gesture. The app will detect landmarks using MediaPipe and classify the ASL letter using YOLOv11.",
+    allow_flagging="never"
+)
+
+if __name__ == "__main__":
+    demo.launch()

yolov11n_finetuned_ASL.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f5f1e19e26f3024d8b7ba2df598bf6981728cdc6ebd133975a85af08e5682c2
+size 3246722