Topics and Questions We Can Help You To Answer:
Paper Instructions:

In the analysis, normal and abdominal findings can be detected at a high rate. A total of 10000 endoscopic images were used and training and testing procedures to be performed AlexNet, GoogleNet and ResNet-50 models. At the same time, training sensitivity of proposed Convolutional Neural Network models was evaluated using different training data numbers. Softmax and Support Vector Machine algorithms were used as classifier. In the proposed method concept, CNN + LSTM model is quite successful compared to the results obtained with other models.A total of 10 class in the data set were obtained with an accuracy rate of 97.90% by proposed method. As a result, analysis of endoscopic images with the proposed algorithms is quite successful.