Some Improvements on Deep Convolutional Neural Network Based Image Classif ication

本文的大概思想就是:

       (1)增加训练样本;(2)增加测试样本预测数量;(3)多个CNN模型的融合;

一、增加训练样本的数量

     常用的增加样本的方法有:crop、flip及add randomly generated ligthing;

    1、传统的crop方法是把图像resize到256*256,然后在进行crop,但是这样会损失掉一部分有用的信息例如下图:

     

      因此本文采用的方法:先将图像的最小的一边放大到256,这样就形成了256*N或者N*256,然后在进行crop;

    2、除了随机加入光照噪声以外,还可以加入额外的颜色处理,即随机改变图像的对比度、亮度和色度(0.5-1.5的比例);

二、增加额外的预测数据

    除了采用五个crops和两个flips外:

    1、采用三个尺度的图像,256、228和284;

    2、采用三个视觉角度,同时倍增了全连的层数,但是并没有取得好的效果;

     

      

    3、减少预测的数量

    由于采用了 translations, 2 flips, 3 scales, and 3 views 这样就产生了90 预测,如果将这90个预测都使用的话,将极大影响测试速度,不使用。简单的方法是采用贪婪方法:即首先采用最可行的一个预测,逐渐加入其他预测,直到识别率不再增加。(不解之处:在测试的时候,根本就无法知道什么时候识别率不再增加,有可能还是要把所有的预测都用上)。效果图如下:

     


     

三、高分辨率的模型

    采用更大的图像,训练一个高分辨率的模型,然后把基本模型和高分辨率模型进行结合。在训练高分辨率的模型的时候有一下几个点:

    1、9 crops, 2 flips, 3 scales, 3 views,采用前面的贪婪预测方法。

    2、由于训练样本非常丰富,因此drop out 就不再那么重要。本文采用分阶段drop out,取得了不使用和一直使用drop out更好的效果。模型融合的效果如下图:

        

四、在ILSVRC2013上的识别率

        

 五、总结

   个人认为,这篇论文也没有什么比较出色的创新点,只不过就是比别人增加样本的方法多一些,并且取得比较好的效果是因为有10个模型的融合的结果。

   值得借鉴的点:

   1、增加样本的方法;

   2、阶段性使用drop out;阶段性使用drop out还属首次,取得了好的效果。

  • 1
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 1
    评论
Deep person re-identification is the task of recognizing a person across different camera views in a surveillance system. It is a challenging problem due to variations in lighting, pose, and occlusion. To address this problem, researchers have proposed various deep learning models that can learn discriminative features for person re-identification. However, achieving state-of-the-art performance often requires carefully designed training strategies and model architectures. One approach to improving the performance of deep person re-identification is to use a "bag of tricks" consisting of various techniques that have been shown to be effective in other computer vision tasks. These techniques include data augmentation, label smoothing, mixup, warm-up learning rates, and more. By combining these techniques, researchers have been able to achieve significant improvements in re-identification accuracy. In addition to using a bag of tricks, it is also important to establish a strong baseline for deep person re-identification. A strong baseline provides a foundation for future research and enables fair comparisons between different methods. A typical baseline for re-identification consists of a deep convolutional neural network (CNN) trained on a large-scale dataset such as Market-1501 or DukeMTMC-reID. The baseline should also include appropriate data preprocessing, such as resizing and normalization, and evaluation metrics, such as mean average precision (mAP) and cumulative matching characteristic (CMC) curves. Overall, combining a bag of tricks with a strong baseline can lead to significant improvements in deep person re-identification performance. This can have important practical applications in surveillance systems, where accurate person recognition is essential for ensuring public safety.
评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值