FICAL: Focal Inter-Class Angular Loss for Image Classification

Wei, Xinran; Chang, Dongliang; Xie, Jiyang; Zheng, Yixiao; Gong, Chen; Zhang, Chuang; Ma, Zhanyu

doi:10.1109/vcip47243.2019.8965889

Cited by 4 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two CNN backbones, VGG16 and ResNet18, are investigated. Furthermore, we compare the CC-Loss with other loss functions including the CE-Loss [12], the Focal loss [23], the A-softmax loss [15], the inter-class angular loss (ICAL) [19], and the focal inter class angular loss (FICAL) [22].…”

Section: Resultsmentioning

confidence: 99%

“…The margin is defined as maximum absolute difference between the five results and it's average. For baseline methods in Table 1 and Table 4, we refer to results from [21] and [22], which only report average values.…”

Section: Implementation Detailsmentioning

confidence: 99%

“…The arcface loss [18] further develops the angular margin to a addictive angular margin, using summation rather than multiplication for efficient training. The inter-class angular loss (ICAL) [19] and the focal inter class angular loss (FICAL) [22] cosine distances between the categories, these two loss functions both obtained higher classification accuracy among these loss functions. In addition to these improvements on loss functions for image classification, the focal loss [23] is a loss function proposed for object detection, assigning bigger weights to hard examples and smaller weights to easy examples to deal with class imbalance problem.…”

Section: Introductionmentioning

confidence: 97%

See 2 more Smart Citations

CC-Loss: Channel Correlation Loss for Image Classification

Song

Chang

et al. 2021

2020 25th International Conference on Pattern Recognition (ICPR)

Self Cite

View full text Add to dashboard Cite

The loss function is a key component in deep learning models. A commonly used loss function for classification is the cross entropy loss, which is a simple yet effective application of information theory for classification problems. Based on this loss, many other loss functions have been proposed, e.g., by adding intra-class and inter-class constraints to enhance the discriminative ability of the learned features. However, these loss functions fail to consider the connections between the feature distribution and the model structure. Aiming at addressing this problem, we propose a channel correlation loss (CC-Loss) that is able to constrain the specific relations between classes and channels as well as maintain the intra-class and the inter-class separability. CC-Loss uses a channel attention module to generate channel attention of features for each sample in the training stage. Next, an Euclidean distance matrix is calculated to make the channel attention vectors associated with the same class become identical and to increase the difference between different classes. Finally, we obtain a feature embedding with good intra-class compactness and inter-class separability. Experimental results show that two different backbone models trained with the proposed CC-Loss outperform the state-of-the-art loss functions on three image classification datasets.

show abstract