It is generally difficult to measure complex shapes such as stairs with high accuracy for indoor environment scanning by the robot. Therefore, we consider the three-dimensional (3D) reconstruction of stairs using Structure from Motion (SfM) and Multi-View Stereo (MVS) which perform 3D reconstruction from images acquired by the robot vision. In this study, we verify whether it is possible to acquire a 3D reconstruction result of stairs by using images shot while ascending and descending the stairs as input to the reconstruction method. To calculate the accuracy of the reconstruction result, we use 3D computer graphics software to generate artificial image data to be applied to the 3D reconstruction. Experimental results show that 3D reconstruction results of the stairs are more accurate by applying both images shot when ascending and descending stairs to the 3D reconstruction methods.

The development and operation of Online Judge System (OJS), which is used to evaluate the correctness of programs, is a nontrivial and difficult task due to the various functional and non-functional requirements. However, although many OJSs have been developed and operated, and their usefulness reported, the theory for constructing OJSs has not been sufficiently discussed. In this paper, we present the functional and nonfunctional requirements oriented to OJS as well as demonstrate the internal components and software architecture of an OJS, which has been in operation for over a decade and has evaluated over six million solutions. We also present real-world experiences and challenges encountered during this long journey of our OJS.

Logical errors in source code can be detected by probabilities obtained from a language model trained by the recurrent neural network (RNN). Using the probabilities and determining thresholds, places that are likely to be logic errors can be enumerated. However, when the threshold is set inappropriately, user may miss true logical errors because of passive extraction or unnecessary elements obtained from excessive extraction. Moreover, the probabilities of output from the language model are different for each task, so the threshold should be selected properly. In this paper, we propose a logic error detection algorithm using an RNN and an automatic threshold determination method. The proposed method selects thresholds using incorrect codes and can enhance the detection performance of the trained language model. For evaluating the proposed method, experiments with data from an online judge system, which is one of the educational systems that provide the automated judge for many programming tasks, are conducted. The experimental results show that the selected thresholds can be used to improve the logic error detection performance of the trained language model.