Class-specific early exit design methodology for convolutional neural networks

Purpose The purpose of this paper is to use convolutional neural networks in order to solve the problem of the difficulty in the classification of cashmere and wool. To do the research, it proposes a low-dimensional strategy of using part-level features to enhance object-level features. The study aims to use computer version method to find out the most effective and robust method to manage the difficult task of cashmere and wool identification. Design/methodology/approach The authors try to get a coarse classification result and the initial weights of the model in the first step. The authors use the results of the first step and a Fast-RCNN method to extract part-level features in step 2. Finally, the authors mix the part-level features to enhance object-level features and classify the cashmere and wool images. Findings The paper finds that not only the texture is the key element of the cashmere and wool identification but also the image colors. Originality/value Most importantly, the paper finds that the part-level features can enhance object-level features in the fiber identification task. However, it does not work in contrast, and the strategy can be used in the similar fibers identifications.

Download Full-text

Towards Edge Computing Using Early-Exit Convolutional Neural Networks

Information ◽

10.3390/info12100431 ◽

2021 ◽

Vol 12 (10) ◽

pp. 431

Author(s):

Roberto G. Pacheco ◽

Kaylani Bochie ◽

Mateus S. Gilbert ◽

Rodrigo S. Couto ◽

Miguel Elias M. Campista

Keyword(s):

Neural Networks ◽

Computer Vision ◽

Convolutional Neural Networks ◽

Low Latency ◽

Application Performance ◽

Early Exit ◽

Network Load ◽

Intermediate Layers ◽

Early Classification ◽

Computer Vision Applications

In computer vision applications, mobile devices can transfer the inference of Convolutional Neural Networks (CNNs) to the cloud due to their computational restrictions. Nevertheless, besides introducing more network load concerning the cloud, this approach can make unfeasible applications that require low latency. A possible solution is to use CNNs with early exits at the network edge. These CNNs can pre-classify part of the samples in the intermediate layers based on a confidence criterion. Hence, the device sends to the cloud only samples that have not been satisfactorily classified. This work evaluates the performance of these CNNs at the computational edge, considering an object detection application. For this, we employ a MobiletNetV2 with early exits. The experiments show that the early classification can reduce the data load and the inference time without imposing losses to the application performance.

Download Full-text

Evaluation von Bildparametern und eines convolutional neural networks in der FDG-PET/MR/CT zur Prädiktion des Gesamtüberlebens (OS) und des Therapieansprechens bei Patienten mit Melanom unter CIT.

10.1055/s-0040-1703340 ◽

2020 ◽

Author(s):

F Seith ◽

J Vogel ◽

C la Fougère ◽

T Küstner ◽

K Nikolaou ◽

...

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Fdg Pet

Download Full-text

Traffic Lights Detection in adverse conditions using Convolutional Neural Networks

10.33107/ubt-ic.2018.75 ◽

2018 ◽

Author(s):

George Symeonidis ◽

Peter P. Groumpos ◽

Evangelos Dermatas

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Traffic Lights ◽

Adverse Conditions

Download Full-text

AUTOMATIC DIAGNOSIS OF BREAST CANCER IN HISTOLOGY IMAGES USING DEEP CONVOLUTIONAL NEURAL NETWORKS

KỶ YẾU HỘI NGHỊ KHOA HỌC CÔNG NGHỆ QUỐC GIA LẦN THỨ XI NGHIÊN CỨU CƠ BẢN VÀ ỨNG DỤNG CÔNG NGHỆ THÔNG TIN ◽

10.15625/vap.2018.0009 ◽

2018 ◽

Author(s):

Hung Le Minh ◽

Manh Mai Van ◽

Toan Tran Dinh ◽

Tot Tran Dac ◽

Tran Van Lang

Keyword(s):

Breast Cancer ◽

Neural Networks ◽

Convolutional Neural Networks ◽

Deep Convolutional Neural Networks ◽

Automatic Diagnosis

Download Full-text

CNN-based Classification of Degraded Images

Electronic Imaging ◽

10.2352/issn.2470-1173.2020.10.ipas-028 ◽

2020 ◽

Vol 2020 (10) ◽

pp. 28-1-28-7 ◽

Cited By ~ 1

Author(s):

Kazuki Endo ◽

Masayuki Tanaka ◽

Masatoshi Okutomi

Keyword(s):

Neural Networks ◽

Image Restoration ◽

Image Classification ◽

Convolutional Neural Networks ◽

Deep Convolutional Neural Networks ◽

Alternative Approach ◽

Degraded Image ◽

Degraded Images ◽

Straightforward Approach

Classification of degraded images is very important in practice because images are usually degraded by compression, noise, blurring, etc. Nevertheless, most of the research in image classification only focuses on clean images without any degradation. Some papers have already proposed deep convolutional neural networks composed of an image restoration network and a classification network to classify degraded images. This paper proposes an alternative approach in which we use a degraded image and an additional degradation parameter for classification. The proposed classification network has two inputs which are the degraded image and the degradation parameter. The estimation network of degradation parameters is also incorporated if degradation parameters of degraded images are unknown. The experimental results showed that the proposed method outperforms a straightforward approach where the classification network is trained with degraded images only.

Download Full-text