深度学习项目班-预习课第二节-卷积神经网络与计算机视觉


卷积神经网络与计算机视觉

Recap

Cross-Entropy vs. Hinge

Cross-Entropy vs. Hinge

Compute the Gradient

计算梯度。

Compute the Gradient

反向传播,链式法则。

Partial Gradient

激活函数及其求导。

Activation Functions

Regularization for Neural Networks - Dropout

Dropout 是 DNN 的一种正则化技术。

Dropout

Convolutional

卷积核有三个参数,width, height, depth.

卷积

ConvNets Architecture Overview

  • A simple ConvNet for CIFAR-10 classification could have the architecture [INPUT - CONV - RELU - POOL - FC]. In more detail:
  • INPUT [32x32x3] will hold the raw pixel values of the image, in this case an image of width 32, height 32, and with three color channels R,G,B.
  • CONV layer will compute the output of neurons that are connected to local regions in the input, each computing a dot product between their weights and the region they are connected to in the input volume. This may result in volume such as [32x32x12].
  • RELU layer will apply an elementwise activation function, such as the max(0,x). This leaves the size of the volume unchanged ([32x32x12]).
  • POOL layer will perform a downsampling operation along the spatial dimensions (width, height), resulting in volume such as [16x16x12].
  • FC (i.e. fully-connected) layer will compute the class scores, resulting in volume of size [1x1x10], where each of the 10 numbers correspond to a class score, such as among the 10 categories of CIFAR-10.


Modern ConvNets

  • AlexNet. Winner of ImageNet 2012. It significantly outperformed the second runner-up (top 5 error of 16% compared to runner-up with 26% error).
  • ZF-Net. Winner of ImageNet 2013. It was an improvement on AlexNet by tweaking the architecture hyperparameters.
  • GoogLeNet. Winner of ImageNet 2014. Its main contribution was the development of an Inception Module that dramatically reduced the number of parameters in the network (4M, compared to AlexNet with 60M).
  • VGGNet. Runner-up of ImageNet 2014. Its main contribution was in showing that the depth of the network is a critical component for good performance. A downside of the VGGNet is that it is more expensive to evaluate and uses a lot more memory and parameters (140M).
  • ResNet. Runner-up of ImageNet 2015. It features special skip connections and a heavy use of batch normalization.
  • Inception-V4. Inception+ResNet.


NCFM – Marine Fish Classification

github 项目开源地址:github repo

Using Keras + TensorFlow to solve NCFM-Leadboard Top 5%.

使用先用的模型,使用迁移学习的方式,进行 fine-tuning

finetune

使用 Inception-V3 模型,进行改进。

Inception-V3
已邀请:

要回复问题请先登录注册

返回顶部