CF-CNN: Coarse-to-Fine Convolutional Neural Network

Abstract

In this paper, we present a coarse-to-fine convolutional neural network (CF-CNN) for learning multilabel classes. The basis of the proposed CF-CNN is a disjoint grouping method that first creates a class group with hierarchical association, and then assigns a new label to a class belonging to each group so that each class acquires multiple labels. CF-CNN consists of one main network and two subnetworks. Each subnetwork performs coarse prediction using the group labels created by the disjoint grouping method. The main network includes a refine convolution layer and performs fine prediction to fuse the feature maps acquired from the subnetwork. The generated class set in the upper level has the same classification boundary to that in the lower level. Since the classes belonging to the upper level label are classified with a higher priority, parameter optimization becomes easier. In experimental results, the proposed method is applied to various classification tasks to show a higher classification accuracy by up to 3% with a much smaller number of parameters without modification of the baseline model.

Modification of a Conventional Deep Learning Model to Classify Simulated Breathing Patterns: A Step toward Real-Time Monitoring of Patients with Respiratory Infectious Diseases

“…To compare the classification performance of our proposed method with other advanced deep learning classification algorithms, we used EfficientNetV2 m [ 55 ], PyramidNet [ 51 ], and CF-CNN [ 52 ]. EfficientNetV2 m and PyramidNet models are CNN-based algorithms developed for image classification and are being utilized in various fields [ 56 , 57 , 58 , 59 , 60 ].…”

“…Short connections mitigate gradient loss by skipping one or more layers during backpropagation. This short connection has been applied to various CNN models [ 50 , 51 , 52 ]. The residual unit is defined as follows: where and represent the input and output features of units, respectively.…”

“…The classification of the datasets split on the subject level yields slightly less accurate results compared to the classification of the datasets of all the subjects. To compare the classification performance of our proposed method with other advanced deep learning classification algorithms, we used EfficientNetV2 m [55], Pyramid-Net [51], and CF-CNN [52]. EfficientNetV2 m and PyramidNet models are CNN-based algorithms developed for image classification and are being utilized in various fields [56][57][58][59][60].…”

All Mistakes Are Not Equal: Comprehensive Hierarchy Aware Multi-label Predictions (CHAMP)

“…Hierarchical information is important in many other applications such as food recognition [55], [25], protein function prediction [6], [7], [56], [57], [58], [59] , image annotation [60], text classification [61], [62], [63]. Some major approaches include imposing logical constraints [4], using hyperbolic embeddings [64], prototype learning [14], label smearing and soft labels, loss modifications [3], multiple learning heads for different levels of the hierarchy [5], hierarchical post-processing [65] and others [66], [67], [68].…”

“…Single label hierarchical (SLH) algorithms encode domain knowledge using tree structures with labels to improve classification performance and make better errors by penalizing a misprediction depending on its severity as per the given hierarchy tree. To achieve this, popular ideas are hierarchy aware custom loss functions [3], label embeddings [4] and custom model architectures [5]. Extending these ideas from single-label hierarchical systems to multilabel scenarios is difficult primarily because these ideas are tailormade for multi-class classification loss functions encountered in SLH but not for multiple separate classification problems as encountered in HMC.…”