sparshasaha/hand-gesture-recognition-using-background-elllimination-and-convolution-neural-network

Hand Gesture Recognition using Convolution Neural Network built using Tensorflow, OpenCV and python

hand-gesture-recognition-using-background-elllimination-and-convolution-neural-network's Issues

ValueError: need more than 2 values to unpack

ValueError: need more than 2 values to unpack when running ContinuousGesturePredictor.py

prediction = model.predict([gray_image.reshape(89,100,1)]) ValueError: cannot reshape array of size 7500 into shape (89,100,1)

network structure

Hello, I have a question about the network structure. Why design seven hidden layers? That sounds a bit too much. I calculated that if the input is 89x100, then the output of Layer 6 convolution layer is 1x1, then what about Layer 7? I am a beginner, so the questions are very basic. I hope you can answer them. Thank you.

[Training Phase] cannot reshape array of size 107100 into shape (89,100,1)

While trying to play around with the .ipynb, I get the following error

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-5-a59e65dd974e> in <module>
     17     image = cv2.imread('Dataset/FistTest/fist_' + str(i) + '.png')
     18     gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
---> 19     testImages.append(gray_image.reshape(89, 100, 1))
     20 
     21 testLabels = []

ValueError: cannot reshape array of size 107100 into shape (89,100,1)

The only thing that I'm doing, just to play around, is to modify the network's structure to this

convnet=input_data(shape=[None,89,100,1],name='input')
convnet=conv_2d(convnet,32,2,activation='relu')
convnet=max_pool_2d(convnet,2)
convnet=conv_2d(convnet,64,2,activation='relu')
convnet=max_pool_2d(convnet,2)

#convnet=conv_2d(convnet,128,2,activation='relu')
#convnet=max_pool_2d(convnet,2)

convnet=conv_2d(convnet,256,2,activation='relu')
convnet=max_pool_2d(convnet,2)

convnet=conv_2d(convnet,256,2,activation='relu')
convnet=max_pool_2d(convnet,2)

convnet=conv_2d(convnet,128,2,activation='relu')
convnet=max_pool_2d(convnet,2)

convnet=conv_2d(convnet,64,2,activation='relu')
convnet=max_pool_2d(convnet,2)

convnet=fully_connected(convnet,1000,activation='relu')
convnet=dropout(convnet,0.75)

convnet=fully_connected(convnet,3,activation='softmax')

convnet=regression(convnet,optimizer='adam',learning_rate=0.001,loss='categorical_crossentropy',name='regression')

model=tflearn.DNN(convnet,tensorboard_verbose=0)

But the crashing point is called before the network's inizialization so I can't get, honestly, what's going wrong :/

New Gesture Dataset

I added my custom data set and trained the model also but when i add new lines in ContinuousGesturePredictor.py , it gives an error.

def getPredictedClass():
# Predict
image = cv2.imread('Temp.png')
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
prediction = model.predict([gray_image.reshape(89, 100, 1)])
return np.argmax(prediction), (np.amax(prediction) / (prediction[0][0] + prediction[0][1] + prediction[0][2] + prediction[0][3]))

def showStatistics(predictedClass, confidence):

textImage = np.zeros((300,512,3), np.uint8)
className = ""

if predictedClass == 0:
    className = "Swing"
elif predictedClass == 1:
    className = "Palm"
elif predictedClass == 2:
    className = "Fist"
elif predictedClass == 3:
    className = "One"

Help pls!!

new dataset

i am trying to create dataset for peace sign. But new threshold images are not saved in my PeaceTest folder even after i changed code in line 103

problem sloved

Version?

What version of TensorFlow and tflearn are you using?

Error on name 'camera' is not defined on PalmTracker file

ValueError: The passed save_path is not a valid checkpoint

C:\Users\adeebx1\Documents\GitHub\Hand-Gesture-Recognition-Using-Background-Elllimination-and-Convolution-Neural-Network>ContinuousGesturePredictor.py
curses is not supported on this machine (please install/reinstall curses for an optimal experience)
WARNING:tensorflow:From C:\Users\adeebx1\AppData\Local\Programs\Python\Python36\lib\site-packages\tflearn\initializations.py:119: UniformUnitScaling.init (from tensorflow.python.ops.init_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.initializers.variance_scaling instead with distribution=uniform to get equivalent behavior.
WARNING:tensorflow:From C:\Users\adeebx1\AppData\Local\Programs\Python\Python36\lib\site-packages\tflearn\objectives.py:66: calling reduce_sum (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.
Instructions for updating:
keep_dims is deprecated, use keepdims instead
Traceback (most recent call last):
File "C:\Users\adeebx1\Documents\GitHub\Hand-Gesture-Recognition-Using-Background-Elllimination-and-Convolution-Neural-Network\ContinuousGesturePredictor.py", line 206, in
model.load("GestureRecogModel.tfl")
File "C:\Users\adeebx1\AppData\Local\Programs\Python\Python36\lib\site-packages\tflearn\models\dnn.py", line 308, in load
self.trainer.restore(model_file, weights_only, **optargs)
File "C:\Users\adeebx1\AppData\Local\Programs\Python\Python36\lib\site-packages\tflearn\helpers\trainer.py", line 490, in restore
self.restorer.restore(self.session, model_file)
File "C:\Users\adeebx1\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python\training\saver.py", line 1538, in restore
+ compat.as_text(save_path))
ValueError: The passed save_path is not a valid checkpoint: C:\Users\adeebx1\Documents\GitHub\Hand-Gesture-Recognition-Using-Background-Elllimination-and-Convolution-Neural-Network\GestureRecogModel.tfl

how to solve this issue!!

AttributeError: 'NoneType' object has no attribute 'shape'

AttributeError: 'NoneType' object has no attribute 'shape'.

In ContinuousGesturePredictor, I tried moving 'import cv2' line to the top. And it SOLVED the issue.