原文传递
Optimization of Convolutional Neural Networks for Enhanced Compression Techniques and Computer Vision Applications.
| 题名： | Optimization of Convolutional Neural Networks for Enhanced Compression Techniques and Computer Vision Applications. |
| 作者： | Couture Del Valle, C. J. |
| 摘要： | Image compression algorithms are the basis of media transmission and compression in the field of image processing. Decades after theirinception, algorithms such as the JPEG image codec continue to be the industry standard. A notable research topic gathering momentum in thefield of compression is deep learning (DL). This paper explores the optimization of DL models for ideal image compression and objectdetection (OD) applications. The DL model to be optimized is based upon an existing compression framework known as the CONNECT model.This framework wraps the traditional JPEG image codec within two convolutional neural networks (CNNs). The first network, ComCNN,focuses on compressing an input image into a compact representation to be fed into the image codec. The second network, RecCNN, focuses onreconstructing the output image from the codec as similarly as possible to the original image. To enhance the performance of the CONNECTmodel, an optimization software called Optuna wraps the framework. Hyperparameters are selected from each CNN to be evaluated andoptimized by Optuna. Once the CONNECT model produces ideal results, the output images are applied to the YOLOv5 OD network. Thispaper explores the impact of DL hyperparameters on image quality and compression metrics. In addition, a detection network will providecontext to the effect of image compression on computer vision applications. |
| 总页数： | 85 pages |
相关文献
- Linear Layers and Partial Weight Reinitialization for Accelerating Neural Network Convergence.
- Competitively Learned Attention Mechanism Prototypes for Network Intrusion Detection.
- Transfer Learning in Convolutional Neural Networks for Fine-Grained Image Classification.
- Toward Automated Aerial Refueling: Automated Visual Aircraft Identification with Convolutional Neural Networks.
- Convolutional Neural Networks For Detection And Classification Of Maritime Vessels In Electro-Optical Satellite Imagery.
- Semantic Segmentation for Aerial Imagery Using U-Nets.
检索历史
应用推荐
