Upload 3 files

app.py

@@ -0,0 +1,64 @@
+import streamlit as st
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+
+# === CONFIG ===
+MODEL_PATH = 'trained_model/asl_model.h5'
+IMG_SIZE = 64
+CLASS_NAMES = [chr(i) for i in range(65, 91)]  # A-Z
+
+# Load model once
+@st.cache_resource(show_spinner=False)
+def load_model():
+    return tf.keras.models.load_model(MODEL_PATH)
+
+model = load_model()
+
+# === UI Header ===
+st.set_page_config(page_title="ASL Recognition", page_icon="🧠", layout="centered")
+st.markdown("<h1 style='text-align: center;'>🧠 ASL Alphabet Recognition</h1>", unsafe_allow_html=True)
+st.markdown("<p style='text-align: center;'>Upload a hand gesture image and get instant letter prediction.</p>", unsafe_allow_html=True)
+st.divider()
+
+# === Helper Functions ===
+def preprocess_image(image: Image.Image):
+    img = image.convert("RGB")
+    img = img.resize((IMG_SIZE, IMG_SIZE))
+    img = np.array(img) / 255.0
+    img = np.expand_dims(img, axis=0)
+    return img
+
+def predict(img: Image.Image):
+    processed = preprocess_image(img)
+    preds = model.predict(processed)
+    class_idx = np.argmax(preds)
+    confidence = preds[0][class_idx]
+    return CLASS_NAMES[class_idx], confidence
+
+# === Upload UI ===
+uploaded_file = st.file_uploader("📁 Upload a hand gesture image", type=['png', 'jpg', 'jpeg'])
+
+if uploaded_file:
+    col1, col2 = st.columns([1, 2])
+    with col1:
+        img = Image.open(uploaded_file)
+        st.image(img, caption="📷 Uploaded Image", use_column_width=True)
+    with col2:
+        st.write("### 🔍 Prediction")
+        label, confidence = predict(img)
+        st.success(f"Predicted Letter: **:blue[{label}]**")
+        st.metric(label="Confidence Score", value=f"{confidence * 100:.2f}%", delta=None)
+
+        # Optional: show full probabilities as a horizontal bar chart
+        preds = model.predict(preprocess_image(img))[0]
+        top_indices = np.argsort(preds)[::-1][:5]
+        st.write("#### 🔢 Top 5 Predictions")
+        for i in top_indices:
+            st.progress(float(preds[i]), text=f"{CLASS_NAMES[i]}: {preds[i]*100:.2f}%")
+else:
+    st.info("📸 Upload a clear image showing a single hand gesture on a plain background.")
+
+# === Footer ===
+st.divider()
+st.markdown("<small style='text-align:center; display:block;'>Developed with ❤️ using TensorFlow & Streamlit</small>", unsafe_allow_html=True)

requirements.txt

@@ -1,3 +1,5 @@
-altair
-pandas
-streamlit
+tensorflow
+streamlit
+opencv-python-headless
+numpy
+matplotlib

train_model.py

@@ -0,0 +1,104 @@
+import os
+import numpy as np
+import tensorflow as tf
+from tensorflow.keras.preprocessing.image import ImageDataGenerator
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
+from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint
+import matplotlib.pyplot as plt
+
+# === CONFIGURATION ===
+DATA_DIR = 'asl_alphabet_train'   # Folder with A-Z subfolders containing images
+MODEL_SAVE_PATH = 'trained_model/asl_model.h5'
+IMG_SIZE = 64
+BATCH_SIZE = 32
+EPOCHS = 20
+NUM_CLASSES = 26
+
+# Create output directories if they don't exist
+os.makedirs(os.path.dirname(MODEL_SAVE_PATH), exist_ok=True)
+os.makedirs('outputs', exist_ok=True)
+
+# === DATA GENERATORS ===
+train_datagen = ImageDataGenerator(
+    rescale=1./255,
+    validation_split=0.2,
+    rotation_range=15,
+    zoom_range=0.1,
+    width_shift_range=0.1,
+    height_shift_range=0.1,
+    horizontal_flip=True
+)
+
+train_generator = train_datagen.flow_from_directory(
+    DATA_DIR,
+    target_size=(IMG_SIZE, IMG_SIZE),
+    batch_size=BATCH_SIZE,
+    class_mode='categorical',
+    subset='training',
+    shuffle=True,
+    seed=42
+)
+
+validation_generator = train_datagen.flow_from_directory(
+    DATA_DIR,
+    target_size=(IMG_SIZE, IMG_SIZE),
+    batch_size=BATCH_SIZE,
+    class_mode='categorical',
+    subset='validation',
+    shuffle=False,
+    seed=42
+)
+
+# === MODEL ARCHITECTURE ===
+model = Sequential([
+    Conv2D(32, (3,3), activation='relu', input_shape=(IMG_SIZE, IMG_SIZE, 3)),
+    MaxPooling2D(2,2),
+
+    Conv2D(64, (3,3), activation='relu'),
+    MaxPooling2D(2,2),
+
+    Conv2D(128, (3,3), activation='relu'),
+    MaxPooling2D(2,2),
+
+    Flatten(),
+    Dense(128, activation='relu'),
+    Dropout(0.5),
+    Dense(NUM_CLASSES, activation='softmax')
+])
+
+model.compile(optimizer='adam',
+              loss='categorical_crossentropy',
+              metrics=['accuracy'])
+
+model.summary()
+
+# === CALLBACKS ===
+checkpoint = ModelCheckpoint(MODEL_SAVE_PATH, save_best_only=True, monitor='val_accuracy', mode='max')
+early_stop = EarlyStopping(monitor='val_loss', patience=5, restore_best_weights=True)
+
+# === TRAINING ===
+history = model.fit(
+    train_generator,
+    validation_data=validation_generator,
+    epochs=EPOCHS,
+    callbacks=[checkpoint, early_stop]
+)
+
+# === PLOT TRAINING HISTORY ===
+plt.figure(figsize=(12,5))
+
+plt.subplot(1,2,1)
+plt.plot(history.history['accuracy'], label='Train Accuracy')
+plt.plot(history.history['val_accuracy'], label='Val Accuracy')
+plt.legend()
+plt.title('Accuracy')
+
+plt.subplot(1,2,2)
+plt.plot(history.history['loss'], label='Train Loss')
+plt.plot(history.history['val_loss'], label='Val Loss')
+plt.legend()
+plt.title('Loss')
+
+plt.savefig('outputs/training_plot.png')
+plt.show()