added 56 models

sign-language/main.py

@@ -34,22 +34,38 @@ sign = {
 23:"y"
 }
 
+lastAns , consecutive = "",0
+
 while(True):
     # Capture frame-by-frame
     ret, frame = cap.read()
-    cv2.rectangle(frame,(100,100),(300,300),(0,255,0),0)
-    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
-    roi = gray[100:300, 100:300]
-    resized = cv2.resize(roi, (28,28), interpolation = cv2.INTER_AREA)
-    resized = resized.reshape(-1,28,28,1).astype(float)
-    print(sign[np.argmax(model.predict(resized))])
-    # Our operations on the frame come here
-
-    # Display the resulting frame
-    cv2.imshow('frame',gray)
-
-    if cv2.waitKey(1) & 0xFF == ord('q'):
-        break
+    frame = cv2.flip(frame,1)
+
+    if ret:
+        cv2.rectangle(frame,(100,100),(300,300),(0,255,0),0)
+        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
+        roi = gray[100:300, 100:300]
+        resized = cv2.resize(roi, (28,28), interpolation = cv2.INTER_AREA)
+        image = resized.reshape(-1,28,28,1).astype(float)
+
+        a = cv2.waitKey(1) # waits to see if `esc` is pressed
+        softmax_output = model.predict(image)
+        if np.max(softmax_output) < 0.5:
+            continue
+        res = np.argmax(softmax_output)
+        lastAns = res
+        if res == lastAns:
+            consecutive+=1
+        else:
+            consecutive = 0
+
+        if consecutive == 10:
+            consecutive = 0
+            cv2.putText(frame, sign[res], (100,400), cv2.FONT_HERSHEY_SIMPLEX, 4, (255,255,255), 4)
+            cv2.imshow("new",roi)
+        cv2.imshow("frame", frame)
+        if a == 27: # when `esc` is pressed
+            break
 
 # When everything done, release the capture
 cap.release()

sign-language/prediction.py

 import pandas as pd
 import numpy as np
 import tensorflow as tf
-from keras.models import Sequential,model_from_json,load_model
+from keras.models import Sequential,load_model
 from keras.layers import Dense, Conv2D, Flatten, MaxPooling2D, Dropout
 from sklearn.preprocessing import LabelBinarizer
 
-trainData = pd.read_csv("/content/drive/My Drive/Colab Notebooks/data/sign-language-mnist/sign_mnist_train.csv").values
-testData = pd.read_csv("/content/drive/My Drive/Colab Notebooks/data/sign-language-mnist/sign_mnist_test.csv").values
+trainData = pd.read_csv("data/sign_mnist_train.csv").values
+testData = pd.read_csv("data/sign_mnist_test.csv").values
 
 trainX,trainY = trainData[:,1:],LabelBinarizer().fit_transform(trainData[:,0:1])
 testX,testY = testData[:,1:],LabelBinarizer().fit_transform(testData[:,0:1])
@@ -30,7 +30,7 @@ model.add(Dense(128, activation = 'relu'))
 model.add(Dropout(0.35))
 model.add(Dense(24, activation = 'softmax'))
 
-model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
+model.compile(loss='logcosh', optimizer='adamax', metrics=['accuracy'])
 
 trainX = trainX.reshape(trainX.shape[0], 28, 28, 1)
 testX = testX.reshape(testX.shape[0], 28, 28, 1)
@@ -39,6 +39,6 @@ model.fit(trainX, trainY, validation_data=(testX, testY), epochs=40, batch_size=
 
 scores = model.evaluate(testX, testY, verbose=0)
 
-model.save("/content/drive/My Drive/Colab Notebooks/data/model.h5")
+model.save("models/model.h5")
 
 print("CNN Error: %.2f%%" % (100-scores[1]*100))
\ No newline at end of file