Heuristic Method for Minimizing Model Size of CNN by Combining Multiple Pruning Techniques.

Network pruning techniques have been widely used for compressing computational and memory intensive deep learning models through removing redundant components of the model. According to the pruning granularity, network pruning can be categorized into structured and unstructured methods. The structured pruning removes the large components in a model such as channels or layers, which might reduce the accuracy. The unstructured pruning directly removes mainly the parameters in a model as well as the redundant channels or layers, which might result in an inadequate pruning. To address the limitations of the pruning methods, this paper proposes a heuristic method for minimizing model size. This paper implements an algorithm to combine both the structured and the unstructured pruning methods while maintaining the target accuracy that is configured by its application. We use network slimming for the structured pruning method and deep compression for the unstructured one. Our method achieves a higher compression ratio than the case when the individual pruning method is applied. To show the effectiveness of our proposed method, this paper evaluates our proposed method with actual state-of-the-art CNN models of VGGNet, ResNet and DenseNet under the CIFAR-10 dataset. This paper discusses the performance of the proposed method with the cases of individual usage of the structured and unstructured pruning methods and then proves that our method achieves better performance with higher compression ratio. In the best case of the VGGNet, our method results in a 13× reduction ratio in the model size, and also gives a 15× reduction ratio regarding the pruning time compared with the brute-force search method.

Statistical Extraction Method for Revealing Key Factors from Posture before Initiating Successful Throwing Technique in Judo.

Many methods such as biomechanics and coaching have been proposed to help people learn a certain movement. There have been proposals for methods to discover characteristics of movement based on information obtained from videos and sensors. Especially in sports, it is expected that these methods can provide hints to improve movement skills. However, conventional methods focus on individual movements, and do not consider cases where external factors influence the movement, such as combat sports. In this paper, we propose a novel method called the Extraction for Successful Movement method (XSM method). Applying the method, this paper focuses on throwing techniques in judo to discover key factors that induce successful throwing from the postures right before initiating the throwing techniques. We define candidate factors by observing the video scenes where the throwing techniques are successfully performed. The method demonstrates the significance of the key factors according to the predominance of factors by χ2 test and residual analysis. Applying the XSM method to the dataset obtained from the videos of the Judo World Championships, we demonstrate the validity of the method with discussing the key factors related to the successful throwing techniques.

Stream-Based Visually Lossless Data Compression Applying Variable Bit-Length ADPCM Encoding.

Video applications have become one of the major services in the engineering field, which are implemented by server-client systems connected via the Internet, broadcasting services for mobile devices such as smartphones and surveillance cameras for security. Recently, the majority of video encoding mechanisms to reduce the data rate are mainly lossy compression methods such as the MPEG format. However, when we consider special needs for high-speed communication such as display applications and object detection ones with high accuracy from the video stream, we need to address the encoding mechanism without any loss of pixel information, called visually lossless compression. This paper focuses on the Adaptive Differential Pulse Code Modulation (ADPCM) that encodes a data stream into a constant bit length per data element. However, the conventional ADPCM does not have any mechanism to control dynamically the encoding bit length. We propose a novel ADPCM that provides a mechanism with a variable bit-length control, called ADPCM-VBL, for the encoding/decoding mechanism. Furthermore, since we expect that the encoded data from ADPCM maintains low entropy, we expect to reduce the amount of data by applying a lossless data compression. Applying ADPCM-VBL and a lossless data compression, this paper proposes a video transfer system that controls throughput autonomously in the communication data path. Through evaluations focusing on the aspects of the encoding performance and the image quality, we confirm that the proposed mechanisms effectively work on the applications that needs visually lossless compression by encoding video stream in low latency.

Non-invasive intravital observation of lingual surface features using sliding oral mucoscopy techniques in clinically healthy subjects.

To investigate intravital morphological features of the broader area of the lingual mucosa in clinically healthy subjects, and to attempt to evaluate subclinical conditions, we evaluated detailed intravital morphological features of the lingual mucosa using our newly developed oral contact mucoscopy techniques. Clinically healthy subjects (female: 19-22 years, average age: 20.27 years, and n = 28) were enrolled. A position indicator stain was placed on the lingual mucosal surface, and sliding images were captured and then reconstructed. In addition, the lingual mucosa was divided into six areas, and morphometry of the fungiform and filiform papillae was performed. The results were statistically analyzed. There were two morphological features among clinically healthy subjects involving the filiform papillae: the length of the papillae and the degree of biofilm (tongue coat) deposition. We defined a modified tongue coat index (mTCI) with scores ranging from 0 (tongue coating not visible) to 0.5, 1, 1.5, and 2 (thick tongue coating) for six sections of the tongue dorsum. No subjects received a score of 2. Significant differences were found in the mTCI between the six sections of the tongue dorsum, especially between the posterior areas and the lingual apex. The fungiform papillae of some subjects exhibited elongated morphological changes. Our findings suggest that magnified lingual dorsum examination of a broader area is especially important in accurate screening for subclinical or transient conditions of potential lingual mucosal diseases. For this purpose, our new oral mucoscopy and non-invasive intravital observational techniques were especially effective.