Mind the Pad--CNNs Can Develop Blind Spots

B Alsallakh, N Kokhlikyan, V Miglani, J Yuan… - arxiv preprint arxiv …, 2020‏ - arxiv.org
We show how feature maps in convolutional networks are susceptible to spatial bias. Due to
a combination of architectural choices, the activation at certain locations is systematically …

Svgformer: Representation learning for continuous vector graphics using transformers

D Cao, Z Wang, J Echevarria… - Proceedings of the IEEE …, 2023‏ - openaccess.thecvf.com
Advances in representation learning have led to great success in understanding and
generating data in various domains. However, in modeling vector graphics data, the pure …

A survey of long‐tail item recommendation methods

J Qin - Wireless Communications and Mobile Computing, 2021‏ - Wiley Online Library
Recommender systems represent a critical field of AI technology applications. The core
function of a recommender system is to recommend items of interest to users, but if it is only …

Fingerprint feature extraction by combining texture, minutiae, and frequency spectrum using multi-task CNN

A Takahashi, Y Koda, K Ito… - 2020 IEEE international …, 2020‏ - ieeexplore.ieee.org
Although most fingerprint matching methods utilize minutia points and/or texture of
fingerprint images as fingerprint features, the frequency spectrum is also a useful feature …

[HTML][HTML] Automatic diabetic retinopathy grading via self-knowledge distillation

L Luo, D Xue, X Feng - Electronics, 2020‏ - mdpi.com
Diabetic retinopathy (DR) is a common fundus disease that leads to irreversible blindness,
which plagues the working-age population. Automatic medical imaging diagnosis provides a …

Hdsuper: High-quality and high computational utilization edge super-resolution accelerator with hardware-algorithm co-design techniques

X Zhao, L Chang, D Fan, Z Hu, T Yue… - IEEE Transactions on …, 2024‏ - ieeexplore.ieee.org
Super-resolution (SR) techniques have been employed to construct high-definition images
from low-quality images. Various neural networks have demonstrated excellent image …

MiniExpNet: A small and effective facial expression recognition network based on facial local regions

X **, Z ** - Neurocomputing, 2021‏ - Elsevier
Abstract Deep networks based Facial Expression Recognition (FER) have shown excellent
performance. Due to high computational cost and memory resources, it is hard to develop …

Early neoplasia identification in Barrett's esophagus via attentive hierarchical aggregation and self-distillation

W Hou, L Wang, S Cai, Z Lin, R Yu, J Qin - Medical image analysis, 2021‏ - Elsevier
Automatic surveillance of early neoplasia in Barrett's esophagus (BE) is of great significance
for improving the survival rate of esophageal cancer. It remains, however, a challenging task …

Learning multi-level representations for affective image recognition

H Zhang, D Xu, G Luo, K He - Neural Computing and Applications, 2022‏ - Springer
Images can convey intense affective experiences and affect people on an affective level.
With the prevalence of online pictures and videos, evaluating emotions from visual content …

Noah: Learning pairwise object category attentions for image classification

C Li, A Zhou, A Yao - arxiv preprint arxiv:2402.02377, 2024‏ - arxiv.org
A modern deep neural network (DNN) for image classification tasks typically consists of two
parts: a backbone for feature extraction, and a head for feature encoding and class …