P Lenz

Today, visual recognition systems are still rarely employed in robotics applications. Perhaps one of the main reasons for this is the lack of demanding benchmarks that mimic such scenarios. In this paper, we take advantage of our autonomous driving platform to develop novel challenging benchmarks for the tasks of stereo, optical flow, visual odometry/SLAM and 3D object detection. Our recording platform is equipped with four high resolution video cameras, a Velodyne laser scanner and a state-of-the-art localization system. Our benchmarks comprise 389 stereo and optical flow image pairs, stereo visual odometry sequences of 39.2 km length, and more than 200k 3D object annotations captured in cluttered scenarios (up to 15 cars and 30 pedestrians are visible per image). Results from state-of-the-art algorithms reveal that methods ranking high on established datasets such as Middlebury perform below average when being moved outside the laboratory to the real world. Our goal is to reduce this bias by providing challenging benchmarks with novel difficulties to the computer vision community. Our benchmarks are available online at: www.cvlibs.net/datasets/kitti.

BACKGROUND AND STUDY AIMS: Double-balloon enteroscopy (DBE) is the first choice endoscopic technique for small-bowel visualization. However, preparation and handling of the double-balloon enteroscope is complex. Recently, a single-balloon enteroscopy (SBE) system has been introduced as being a simplified, less-complex balloon-assisted enteroscopy system. PATIENTS AND METHODS: This study was a randomized international multicenter trial comparing two balloon-assisted enteroscopy systems: DBE vs. SBE. Consecutive patients referred for balloon-assisted enteroscopy were randomized to either DBE or SBE. Patients were blinded with regard to the type of instrument used. The primary study outcome was oral insertion depth. Secondary outcomes included complete small-bowel visualization, anal insertion depth, patient discomfort, and adverse events. Patient discomfort during and after the procedure was scored using a visual analog scale. RESULTS: A total of 130 patients were included over 12 months: 65 with DBE and 65 with the SBE technique. Patient and procedure characteristics were comparable between the two groups. Mean oral intubation depth was 253 cm with DBE and 258 cm with SBE, showing noninferiority of SBE vs. DBE. Complete visualization of the small bowel was achieved in 18 % and 11 % of procedures in the DBE and SBE groups, respectively. Mean anal intubation depth was 107 cm in the DBE group and 118 cm in the SBE group. Diagnostic yield and mean pain scores during and after the procedures were similar in the two groups. No adverse events were observed during or after the examinations. CONCLUSIONS: This head-to-head comparison study shows that DBE and SBE have a comparable performance and diagnostic yield for evaluation of the small bowel.

BACKGROUND AND STUDY AIMS: Double-balloon enteroscopy (DBE) has been proven effective for deep intubation of the small bowel. However, intubation depth is limited by distention of the small bowel due to air insufflation during the procedure. The present trial investigated whether carbon dioxide (CO (2)) instead of standard air insufflation would improve intubation depth during DBE, as well as reduce postprocedure pain. PATIENTS AND METHODS: One hundred and twelve consecutive patients scheduled for DBE at two centers were randomly assigned to either CO (2) or air insufflation during DBE. Patients and endoscopists were blinded with regard to the type of gas used. Intubation depth was registered using a validated form. Patients scored pain and discomfort during and after the examination on a 100-mm visual analog scale. RESULTS: One hundred patients were eligible for data analysis (48 in the CO (2) group and 52 in the air group). The mean small-bowel intubation depth was extended by 30 % in the CO (2) group compared to the air group (230 vs. 177 cm, P = 0.008). The superiority was most pronounced for oral DBE, with a 71-cm improvement in intubation depth when using CO (2) (295 cm in the CO (2) group vs. 224 cm in the air group, P < 0.001). Patient pain and discomfort were significantly reduced in the CO (2) group at 1 and 3 hours after the examination. CONCLUSIONS: CO (2) insufflation significantly extended intubation depth in DBE. CO (2) insufflation also reduces patient discomfort. CO (2) insufflation may lead to a higher diagnostic and therapeutic yield of DBE, with reduced patient discomfort.