Light field cameras capture both the spatial and the angular properties of light rays in space. Due to its property, one can compute the depth from light fields in uncontrolled lighting environments, which is a big advantage over active sensing devices. Depth computed from light fields can be used for many applications including 3D modelling and refocusing. However, light field images from hand-held cameras have very narrow baselines with noise, making the depth estimation difficult. Many approaches have been proposed to overcome these limitations for the light field depth estimation, but there is a clear trade-off between the accuracy and the speed in these methods. In this paper, we introduce a fast and accurate light field depth estimation method based on a fully-convolutional neural network. Our network is designed by considering the light field geometry and we also overcome the lack of training data by proposing light field specific data augmentation methods. We achieved the top rank in the HCI 4D Light Field Benchmark on most metrics, and we also demonstrate the effectiveness of the proposed method on real-world light-field images.

PurposeThis study aimed to analyze the learning curves for colorectal surgery fellows in a colonoscopy training program.MethodsBetween May 2003 and February 2017, 60 surgical fellows joined our 1-year colonoscopy training program as trainees and performed 43,784 cases of colonoscopy. All trainees recorded their colonoscopy experiences prospectively into the database. After excluding 6 trainees, who had experience with performing more than 50 colonoscopies before participating in our training program or who discontinued our training program with experience performing less than 300 colonoscopies, this study included 54 trainees who had performed 39,539 colonoscopy cases. We analyzed the cecal intubation rate (CIR) and cecal intubation time (CIT) using the cumulative sum (Cusum) technique and moving average method to assess the technical colonoscopy competence.ResultsOverall, the CIR by the trainees was 80.7%. The median number of cases of colonoscopy performed during the training period for each trainee was 696 (range, 322–1,669). The trainees were able to achieve a 90% CIR with 412 and 493 procedures when analyzed using the moving average and the Cusum, respectively. Using the moving average method, CIRs after 150, 300, and 400 procedures were 67.0%, 84.1%, and 89.2%, respectively. The CIT of trainees continuously decreased until 400 successful cases. Median CITs were 9.4, 8.3, and 7.4 minutes at 150, 300, and 400 successful cases, respectively.ConclusionWe found that more than 400 cases of experience were needed for technical competence in colonoscopy. Continuous teaching and monitoring is required until trainees become sufficiently competent.

No abstract