train_basic_ml.py

# Classical ML Emotion Detection with HOG, LBP, Gabor (NumPy 2.x Compatible - OpenCV Removed)

import numpy as np
from skimage.io import imread
from skimage.transform import resize
from skimage.feature import hog, local_binary_pattern
from skimage.filters import gabor
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report
import os
from tqdm import tqdm
import joblib  # Import joblib for saving models

# ----------------------
# Configuration
# ----------------------
IMG_SIZE = 48
DATASET_PATH = "./train"  # Folder structure: ./dataset/<label>/<image>.jpg
EMOTIONS = ['angry', 'disgust', 'fear', 'happy', 'neutral', 'sad', 'surprise']

# ----------------------
# Feature Extraction Functions
# ----------------------
def extract_hog(img):
    return hog(img, pixels_per_cell=(8, 8), cells_per_block=(2, 2))

def extract_lbp(img):
    lbp = local_binary_pattern(img, P=8, R=1, method="uniform")
    (hist, _) = np.histogram(lbp.ravel(), bins=np.arange(0, 59))
    hist = hist.astype("float")
    hist /= (hist.sum() + 1e-6)
    return hist

def extract_gabor(img):
    filt_real, _ = gabor(img, frequency=0.6)
    return filt_real.ravel()[::64]  # downsample to reduce dimensionality

# ----------------------
# Load Dataset and Extract Features
# ----------------------
def load_dataset():
    data = []
    labels = []
    print("[INFO] Loading dataset and extracting features...")

    for label in EMOTIONS:
        folder = os.path.join(DATASET_PATH, label)
        if not os.path.exists(folder):
            continue
        for file in tqdm(os.listdir(folder), desc=label):
            path = os.path.join(folder, file)
            try:
                img = imread(path, as_gray=True)
                img = resize(img, (IMG_SIZE, IMG_SIZE), anti_aliasing=True)
            except Exception as e:
                print(f"[WARNING] Skipped {file}: {e}")
                continue

            feat = []
            feat.extend(extract_hog(img))
            feat.extend(extract_lbp(img))
            feat.extend(extract_gabor(img))

            data.append(feat)
            labels.append(EMOTIONS.index(label))

    return np.array(data), np.array(labels)

# ----------------------
# Train & Evaluate Models
# ----------------------
def train_and_evaluate(X, y):
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

    classifiers = {
        "SVM": SVC(kernel='linear'),
        "Random Forest": RandomForestClassifier(n_estimators=100),
        "k-NN": KNeighborsClassifier(n_neighbors=5),
        "Logistic Regression": LogisticRegression(max_iter=1000)
    }

    for name, clf in classifiers.items():
        print(f"\n[INFO] Training {name}...")
        clf.fit(X_train, y_train)
        preds = clf.predict(X_test)
        print(f"\n[RESULTS] {name} Classification Report")
        print(classification_report(y_test, preds, target_names=EMOTIONS))
        # Save the trained model using joblib
        model_filename = f'{name.lower().replace(" ", "_")}_model.joblib'
        print(f"[INFO] Saving {name} model to {model_filename}...")
        joblib.dump(clf, model_filename)

# ----------------------
# Run Full Pipeline
# ----------------------
if __name__ == "__main__":
    X, y = load_dataset()
    train_and_evaluate(X, y)