With the aim of achieving communication that reaches the heart, researchers at NTT Communication Science Laboratories have been exploring the essence of communication since its founding 30 years ago. They have been creating technologies that approach and exceed human abilities in such fields as media processing and data science. They have also been discovering basic principles that lead to a deeper understanding of humans in fields such as cognitive neuroscience and brain science. This article introduces some of the laboratories’ activities in pursuit of the essence of communication.

A Cross-Entropy Loss is often used as loss function in training a semantic segmentation network for facial parts. With the Cross-Entropy Loss, the network learns so that the segmentation accuracy of the facial parts having a large number of pixels is high. Therefore, the trained network would not be able to segment face parts with a small number of pixels accurately. In addition, it may be possible to flexibly deal with facial parts of various sizes by comprehensively analyzing features at different resolutions. The purpose of this study was to develop a semantic segmentation method for facial parts using a modified U-Net, which hasMultiple Decoders structure for analyzing features at different resolutions, with Generalized Dice Loss for correcting the bias of the number of pixels in each class. Our database consisted of 30,000 face images from CelebA Mask HQ dataset. The proposed network based on U-Net consists of an encoder, five decoders that perform semantic segmentation independently using feature maps extracted from the encoder at different resolutions, and a recognition that integrates the analysis information in those decoders. The mean intersection over union for the proposed method was 0.846, which was greater than those for SegNet (0.711), U-Net (0.803), and a modified SegNet with Encoder-Multiple Decoders (0.805).

In order to produce a holographic reconstructed image in colors, we must study sufficiently the characteristics of the images reproduced not only by red or green color laser lights but also, in particular, by blue color laser light. But in general,if we use, in reproduction of the images, the laser lights with short wave length, the reconstruction comes to be more difficult. For this reason, it seems to be very few to find the results concerning the reconstruction with the use of blue color laser light.It appears to lead us to an unfortunate problem of the restriction on the color region of the recovered holographic images. Our task is to study carefully the characteristics of the recovered images performed by the blue color laser lights of short wave length, and to make much efforts to enlarge the presentable color region of the reproduced images. In this report,we have studied possibility of the reconstruction of holographic images by the use of blue-violet laser lights. As this result,we have found some condition that makes us possible to perform a reproduction of holographic images by the use of blue-violet laser lights. This seems to suggest some possibility of an extension of the color region for the recovered holographic images.

Online learning at universities has spread rapidly in response to the new Coronavirus infection. However, the major problem with online lectures is that it is very difficult to evaluate how concentrated the students are taking the lectures. Therefore, in this study, we propose a method to evaluate objectively the concentration level of students during online lectures using an eye tracking device. The concentration level is scored using eye movement data of the student viewing the online lecture slides and the total score is added up by computing some important parts of each slide. For the experiment, we made the contents for online lectures, the comprehension test to be conducted after online learning, and the questionnaire after the experiment. In this experiment, the correlation coefficient between the total score calculated by tracking the student’s line of sight for each slide and the score obtained from the confirmation test was calculated. As a result, it was shown that the higher the evaluation score of the concentration level, the score of the comprehension test arose, and the effectiveness of the proposed method for objectively evaluating the concentration level of the students was clarified.