We propose an improved content-aware image resizing method that uses deep learning. The proposed method is extended from seam carving, which is another image resizing method. Seam carving uses the energy map from an image. It also removes a seam where the energy is the minimum. We propose a method for creating a deep energy map using an encoder—decoder convolution neural network. A deep energy map preserves important parts or boundaries in an image, without distortion. Furthermore, it has the characteristic that uniform intensity of edges is displayed for all images. Four well-known resizing methods and our proposed method were evaluated in terms of aspect ratio similarity. In such an objective evaluation, the proposed method demonstrated better results, compared to the other four algorithms. Our proposed method can reduce the size of an image without damaging the overall structure or losing important information in the image.

Thermal sensors are robust to conditions that are constraints for visual sensors, such as illumination changes. Based on this effectiveness, studies on hand tracking have enabled the demonstration of higher performance with low computation times. In this paper, we propose a novel hand detection method and tracking framework based on information extracted from thermal images. An adaptive hand detection (AHD) method was designed with five models that use temperature analysis to obtain the region of interest (ROI) of the hand. To improve performance, we introduced an AHD-based automatic tracking-by-detection algorithm using the Kernelized Correlation Filters (KCF) tracker. Finally, we combined the hand detection and tracking algorithms into a single framework, a guidance framework for tracking by detection (GFTD). To verify the performance, we evaluated the accuracy of the hand detection using intersection over union (IOU) and the trajectory using object tracking error (OTE). The proposed GFTD improves the performance in terms of IOU and OTE by 15% and 16.3%, respectively, compared to the conventional methods.

The decision tree is one of the most effective tools for deriving meaningful outcomes from image data acquired from the visual sensors. Owing to its reliability, superior generalization abilities, and easy implementation, the tree model has been widely used in various applications. However, in image classification problems, conventional tree methods use only a few sparse attributes as the splitting criterion. Consequently, they suffer from several drawbacks in terms of performance and environmental sensitivity. To overcome these limitations, this paper introduces a new tree induction algorithm that classifies images on the basis of local area learning. To train our predictive model, we extract a random local area within the image and use it as a feature for classification. In addition, the self-organizing map, which is a clustering technique, is used for node learning. We also adopt a random sampled optimization technique to search for the optimal node. Finally, each trained node stores the weights that represent the training data and class probabilities. Thus, a recursively trained tree classifies the data hierarchically based on the local similarity at each node. The proposed tree is a type of predictive model that offers benefits in terms of image’s semantic energy conservation compared with conventional tree methods. Consequently, it exhibits improved performance under various conditions, such as noise and illumination changes. Moreover, the proposed algorithm can improve the generalization ability owing to its randomness. In addition, it can be easily applied to ensemble techniques. To evaluate the performance of the proposed algorithm, we perform quantitative and qualitative comparisons with various tree-based methods using four image datasets. The results show that our algorithm not only involves a lower classification error than the conventional methods but also exhibits stable performance even under unfavorable conditions such as noise and illumination changes.

Sunjin Yu, Eungyeol Song, and Changyong Yoon. IJFIS 2018;18:146-53. https://doi.org/10.5391/IJFIS.2018.18.2.146

In multi-cat households, monitoring individual cats’ various behaviors is essential for diagnosing their health and ensuring their well-being. This study focuses on the defecation and urination activities of cats, and introduces an adaptive cat identification architecture based on deep learning (DL) and machine learning (ML) methods. The architecture comprises an object detector and a classification module, with the primary focus on the design of the classification component. The DL object detection algorithm, YOLOv4, is used for the cat object detector, with the convolutional neural network, EfficientNetV2, serving as the backbone for our feature extractor in identity classification with several ML classifiers. Additionally, to address changes in cat composition and individual cat appearances in multi-cat households, we propose an adaptive concept drift approach involving retraining the classification module. To support our research, we compile a comprehensive cat body dataset comprising 8934 images of 36 cats. After a rigorous evaluation of different combinations of DL models and classifiers, we find that the support vector machine (SVM) classifier yields the best performance, achieving an impressive identification accuracy of 94.53%. This outstanding result underscores the effectiveness of the system in accurately identifying cats.

Eungyeol Song, Sun-Woong Yoon, Hanbin Son, and Sunjin Yu. IJFIS 2018;18:167-74. https://doi.org/10.5391/IJFIS.2018.18.3.167