Quick Contact

    Python Tutorial
    Python Panda Tutorial
    Python Selenium
    Python Flask Tutorial
    Python Django
    Interview Questions & Answers



    Solve image recognition problems with very high accuracy. It gives less accuracy for complex problems.


    1. It automatically detected essential features without human help.
    2. Weight sharing is done with this.
    3. But CNN don’t encode the position and orientation of the object.
    4. It lacks the ability of spatially invariant to – input data.
    5. Huge (training) data in order to work efficiently.


    1. RNN stores long time information.
    2. RNN can also used with convolution layers to extend effective pixel neighborhood.
    3. But it has exploding as well as (gradient) vanishing.
    4. RNN code training is very difficult.
    5. In case of activation functions like ‘tanh’ and ‘relu’, it can’t process the huge long sequence.


    – This type of model considers variation of multilayer perceptrons and contains one or more convolution layers that can be either entirely connected or pooled.

    – These convolution layers create feature maps that record a region of image that is ultimately broken into rectangles and sent for nonlinear processing.


    – RNN models save the output of processing nodes and feed the result back into model and hence it don’t pass the information in one direction, so this is how, model set to learn to predict the outcome of layer.

    – Each node in RNN model acts as memory cell, continuing the computation and implementation of operation.

    – If network prediction will be incorrect then system self-learns and continues the working toward the correct prediction during back propagation.


    1. CNN neural network is feed forward, it considers many applications in the sector of image recognition as well as object recognition whereas RNN feedback is fundamentally based, and in RNN the output is also dependent on previous layer.
    2. CNN includes 4 layers i.e. activation layer, convolution layer, fully connected layer and pooling layer. These layers extract the features and find the patterns of input image.
    3. CNN mostly support images.
    4. CNN has finite input set.


    1. RNN layer feedback is based on previous layer also. RNN considers input/hidden as well as output layer. Looping is done by hidden layer and it has memory to store previous results.
    2. RNN support the sequential data.
    3. RNN don’t have finite set of input, it considers arbitrary length of input.
    4. RNN find good with time series information.


    Apply now for Advanced Python Training Course

    Copyright 1999- Ducat Creative, All rights reserved.

    Anda bisa mendapatkan server slot online resmi dan terpercaya tentu saja di sini. Sebagai salah satu provider yang menyediakan banyak pilihan permainan.