In this paper, we propose a two-step method to recognize multiple-food images by detecting candidate regions with several methods and classifying them with various kinds of features. In the first step, we detect several candidate regions by fusing outputs of several region detectors including Felzenszwalb's deformable part model (DPM) [1], a circle detector and the JSEG region segmentation. In the second step, we apply a feature-fusion-based food recognition method for bounding boxes of the candidate regions with various kinds of visual features including bag-of-features of SIFT and CSIFT with spatial pyramid (SP-BoF), histogram of oriented gradient (HoG), and Gabor texture features.In the experiments, we estimated ten food candidates for multiple-food images in the descending order of the confidence scores. As results, we have achieved the 55.8% classification rate, which improved the baseline result in case of using only DPM by 14.3 points, for a multiple-food image data set. This demonstrates that the proposed two-step method is effective for recognition of multiple-food images.

The aims of this study were (1) to evaluate changes in muscle activity associated with physiological fatigue and decreased swimming velocity (SV) during 200 m of front crawl swimming, and (2) to examine the relationship between the decreased SV and changes in kinematic or electromyogram parameters. Twenty swimmers participated in a 4 × 50-m swim test. The surface EMG of 11 muscles (7 in the upper limbs and 4 in the lower limbs) was measured and the mean amplitude value (MAV) for one stroke cycle was obtained. The SV and arm angular velocity (AAV) of shoulder flexion during the first (early stroke) and second (late stroke) half of the underwater arm stroke were analyzed using an underwater camera. The AAV, the MAV of flexor carpi ulnaris (FCU), biceps brachii (BB), and triceps brachii during the early stroke, and the MAV of rectus femoris decreased along with a decrease in SV. In contrast, the MAV of the pectoralis major (PM) increased significantly in the final 50 m. The rate of change in MAVs (ΔMAVs) of FCU, BB and latissimus dorsi during the early stroke, and ΔMAV of biceps femoris were significantly correlated with ΔSV and/or ΔAAV. Positive correlations were identified between ΔMAVs of several muscles. However, no negative correlations were observed between ΔMAVs. These results suggest that the decrease in SV was related to decreases in the activities of several muscles that coordinated with each other, and that a compensating strategy occurred between PM and other muscles in the final 50 m.

The aim of this study was to examine whether the intracyclic velocity variation (IVV) was lower in elite swimmers than in beginner swimmers at various velocities, and whether differences may be related to arm coordination. Seven elite and nine beginner male swimmers swam front crawl at four different swimming velocities (maximal velocity, 75%, 85%, and 95% of maximal swimming velocity). The index of arm coordination (IDC) was calculated as the lag time between the propulsive phases of each arm. IVV was determined from the coefficient of variation of horizontal velocity within one stroke cycle. IVV for elite swimmers was significantly lower (26%) than that for beginner swimmers at all swimming velocities (p<0.01, 7.28 1.25% vs. 9.80 1.70%, respectively). In contrast, the IDC was similar between elite and beginner swimmers. These data suggest that IVV is a strong predictor of the skill level for front crawl, and that elite swimmers have techniques to decrease IVV. However, the IDC does not contribute to IVV differences between elite and beginner swimmers.

No abstract