Vacant parking slot detection using image classification has been studied for a long time. Currently, deep neural networks are widely used in this research field, and experts have concentrated on improving their performance. As a result, most experts are not concerned about the features extracted from the images. Thus, no one knows the crucial features of how neural networks determine whether a particular parking slot is full. This study divides the structures of neural networks into feature extraction and classification parts to address these issues. The output of the feature extraction parts is visualized through normalization and grayscale imaging. The visualized feature maps are analyzed to match the feature characteristics and classification results. The results show that a specific region of feature maps is activated if the parking slot is full. In addition, it is verified that different networks whose classification parts are identical extract similar features from parking slot images. This study demonstrates that feature map analyses help us find hidden characteristics of features and understand how neural networks operate. Our findings show a possibility that handcrafted algorithms using the features found by machine learning algorithms can replace neural network-based classification parts.

While the adoption of wireless capsule endoscopy (WCE) has been steadily increasing, its primary application remains limited to observing the small intestine, with relatively less application in the upper gastrointestinal tract. However, there is a growing anticipation that advancements in capsule endoscopy technology will lead to a significant increase in its application in upper gastrointestinal examinations. This study addresses the underexplored domain of landmark identification within the upper gastrointestinal tract using WCE, acknowledging the limited research and public datasets available in this emerging field. To contribute to the future development of WCE for gastroscopy, a novel approach is proposed. Utilizing color transfer techniques, a simulated WCE dataset tailored for the upper gastrointestinal tract is created. Using Euclidean distance measurements, the similarity between this color-transferred dataset and authentic WCE images is verified. Pioneering the exploration of anatomical landmark classification with WCE data, this study integrates similarity evaluation with image preprocessing and deep learning techniques, specifically employing the DenseNet169 model. As a result, utilizing the color-transferred dataset achieves an anatomical landmark classification accuracy exceeding 90% in the upper gastrointestinal tract. Furthermore, the application of sharpen and detail filters demonstrates an increase in classification accuracy from 91.32% to 94.06%.

In this paper, we propose a new optical encryption technique that uses the single random phase mask. In conventional optical encryptions such as double random phase encryption (DRPE), two different random phase masks are required to encrypt the primary data. For decryption, DRPE requires taking the absolute value of the decrypted data because it is complex-valued. In addition, when key information is revealed, the primary data may be reconstructed by attackers. To reduce the number of random phase masks and enhance the security level, in this paper, we propose single random phase encryption (SRPE) with additive white Gaussian noise (AWGN) and volumetric computational reconstruction (VCR) of integral imaging. In our method, even if key information is known, the primary data may not be reconstructed. To enhance the visual quality of the decrypted data by SRPE, multiple observation is utilized. To reconstruct the primary data, we use VCR of integral imaging because it can remove AWGN by average effect. Thus, since the reconstruction depth can be another key piece of information of SRPE, the security level can be enhanced. In addition, it does not require taking the absolute value of the decrypted data for decryption. To verify the validity of our method, we implement the simulation and calculate performance metrics such as peak sidelobe ratio (PSR) and structural similarity (SSIM). In increasing the number of observations, SSIM for the decrypted data can be improved dramatically. Moreover, even if the number of observations is not enough, three-dimensional (3D) data can be decrypted by SRPE at the correct reconstruction depth.

A notable drawback of Double Random Phase Encryption (DRPE), a prominent optical cryptography technique, is its low data efficiency. This is because both the encrypted image and the decryption key are represented as complex numbers. To address this issue, a conventional method was proposed that encrypts two images simultaneously by treating the first image as amplitude and the second image as phase. Nevertheless, processes such as integral imaging, which extract 3D object information from images, utilize vast amounts of imagery, necessitating further enhancements in data efficiency. The objective of this research is to enhance DRPE and improve data efficiency by increasing the number of images that can be processed simultaneously. This paper incorporates the information from a third image into the random phase mask used in conventional methods, enabling the simultaneous processing of three images. It also proposes a method to synthesize two images by extracting their high-order bits and combining them. The combination of this image composition method as a preprocessing step with the proposed DRPE method enables the simultaneous processing of six images. As a result, the proposed method achieves a data efficiency approximately six times that of the basic DRPE and approximately three times that of conventional methods. The quality of the decrypted images was evaluated using PSNR and SSIM, while the encryption strength was assessed in terms of key space, key sensitivity, entropy, and correlation coefficients.