Lossy image compression always faces a tradeoff between rate-distortion performance and compression/decompression speed. With the advent of neural image compression, hardware (GPU) becomes the new vertex in the tradeoff triangle. By resolving the high GPU dependency and improving the low speed of neural models, this paper proposes two non-GPU models that get the best of the three worlds. First, the CPU-friendly Independent Separable Down-Sampling (ISD) and Up-Sampling (ISU) modules are proposed to lighten the network while ensuring a large receptive field. Second, an asymmetric autoencoder architecture is adopted to boost the decoding speed. At last, the Inverse Quantization Residual (IQR) module is proposed to reduce the error caused by quantization. In terms of rate-distortion performance, our network surpasses the state-of-the-art real-time GPU neural compression work at medium and high bit rates. In terms of speed, our model's compression and decompression speeds surpass all other traditional compression methods except JPEG, using only CPUs. In terms of hardware, the proposed models are CPU friendly and perform stably well in a non-GPU environment. The code is publicly available at https://github.com/kengchikengchi/FasiNet.

In the era of big data, web videos are known for their astronomical volume and great difficulties to be understood by computers. Therefore it is challenging to detect and curb tag pollution on social networking platforms. Intuitively, the pollution in the tags of a video can be identified by exploiting the tags of visually similar videos. From this intuition, we develop a semi-supervised approach to estimate the relevance between Internet videos and user-provided labels accurately and detect polluted labels accordingly. To further enhance the accuracy of relevance estimation and pollution detection, we introduce three multi-view multi-label models, which employ the coherence and differences between various similarity relations of videos. Compared with two quintessential multi-view fusion models, the proposed models consistently outperform or achieve comparable performance.

State-of-the-art indoor localization techniques can reach high localization accuracy, but rely on widely deployed infrastructure which may be absent in less developed regions. To achieve infrastructure-free indoor localization, researchers have proposed to use the inertial sensors on the mobile devices to update the users potential locations by walking steps. Such methods are known to accumulate errors and degrade drastically over time. To compensate for it, either a wall-constraint that eliminates possible steps going through the walls, or landmarks correspond to the special user activities, are usually employed. However, these approaches still suffer from slow convergence and cannot detect floor changes automatically without the information from other users. In this paper, we propose LocMe, an indoor localization service based on human locomotion detection and map exploitation. By synergizing both the locomotion-constraint and wall-constraint, LocMe can significantly increase the converging speed of localization and automatically detect the floor changes, with negligible extra complexity. Our field tests show that LocMe can achieve a median localization error of 1.1 m, which is over 68% lower than the localization algorithm with only the wall-constraint in the same test condition.