No abstract

Imagery texts are usually organized as a hierarchy of several visual elements, i.e. characters, words, text lines and text blocks. Among these elements, character is the most basic one for various languages such as Western, Chinese, Japanese, mathematical expression and etc. It is natural and convenient to construct a common text detection engine based on character detectors. However, training character detectors requires a vast of location annotated characters, which are expensive to obtain. Actually, the existing real text datasets are mostly annotated in word or line level. To remedy this dilemma, we propose a weakly supervised framework that can utilize word annotations, either in tight quadrangles or the more loose bounding boxes, for character detector training. When applied in scene text detection, we are thus able to train a robust character detector by exploiting word annotations in the rich large-scale real scene text datasets, e.g. ICDAR15 [19] and . The character detector acts as a key role in the pipeline of our text detection engine. It achieves the state-of-the-art performance on several challenging scene text detection benchmarks. We also demonstrate the flexibility of our pipeline by various scenarios, including deformed text detection and math expression recognition.

No abstract

Elastic distortion of fingerprints is one of the major causes for false non-match. While this problem affects all fingerprint recognition applications, it is especially dangerous in negative recognition applications, such as watchlist and deduplication applications. In such applications, malicious users may purposely distort their fingerprints to evade identification. In this paper, we proposed novel algorithms to detect and rectify skin distortion based on a single fingerprint image. Distortion detection is viewed as a two-class classification problem, for which the registered ridge orientation map and period map of a fingerprint are used as the feature vector and a SVM classifier is trained to perform the classification task. Distortion rectification (or equivalently distortion field estimation) is viewed as a regression problem, where the input is a distorted fingerprint and the output is the distortion field. To solve this problem, a database (called reference database) of various distorted reference fingerprints and corresponding distortion fields is built in the offline stage, and then in the online stage, the nearest neighbor of the input fingerprint is found in the reference database and the corresponding distortion field is used to transform the input fingerprint into a normal one. Promising results have been obtained on three databases containing many distorted fingerprints, namely FVC2004 DB1, Tsinghua Distorted Fingerprint database, and the NIST SD27 latent fingerprint database.

patches have been recommended by physicians for patients with heart abnormalities to correlate their activities with heart signals. [2] In order for long-duration on-skin monitoring, comfort to the wearer is an important design consideration for these patches. To enable intimate attachment to the body, the package modulus and form factor should approach that of the human skin. Besides enhancing comfort, good electrical contact for high fidelity signal acquisition that is immune to the motion of wearer and environment influences are necessary. [8] Despite significant advancement in wearable electronics, [1, major tradeoffs between form-factor, performance, and functionality remain. For ultrathin skin-like material systems fabricated by ink-printing or microcontact transfer printing, [9][10][11][12][13][14][15][20][21][22][25][26][27] the complexity and signal processing capabilities are typically limited by the weaker transistors or interconnects. Ink-printed components are also limited by lower integration density compared to rigid Silicon CMOS technologies, leading to lower functionality. The increased R-C parasitic with larger and weaker components limits the scalability toward highly-energy-efficiency under low-voltage operations. In contrast, rigid CMOS chips and printed circuit boards (PCBs) require integration with soft components to interface comfortably with the human body. The need for electrical performance with soft and robust mechanical form factor leads us to the codesign of composite materials and electronic circuits/system in a monolithic form of flexible hybrid electronics. [1,3,4,12,13,25,28] In this work, we report on a novel integration of a wearable and stretchable-hybrid SEP with monolithically integrated sensor electrodes and liquid-metal interconnects. The SEP integration involves the combination of a chip-on-board embedded in a moisture-resistant elastomer matrix with microfluidic interconnects, and soft low-resistance electrodes (Figure 1). A stretchable electrocardiogram (ECG) patch (SEP) that monolithically integrates ECG monitoring chip-on-board (COB) with polydimethylsiloxane (PDMS) and liquid-metal interconnects is presented. The 4.8 × 4.8 cm 2 SEPis conformal and robust to mechanical deformation. The use of a siliconon-insulator rigid complementary-metal-oxide-semiconductor chip allows sophisticated power management and signal processing. The chip's dense inputs/output pads are interfaced with coarser liquid-metal interconnects using a dual-sided COB design. A robust ECG signal response (≈100 mV p-p up to 1 kHz), subjected to mechanical deformation and moisture is demonstrated. The SEP allows up to 10% stretch, providing sufficient pliability to enable conformal contact to the human chest. Low profile soft carbon black-PDMS nanocomposite electrodes, robust to deformation, enable good skin contact and allow for low-noise signal acquisition that is comparable to larger commercial wet electrodes.