Modern smartphone cameras can match traditional DSLR cameras in many areas thanks to the introduction of camera arrays and multi-frame processing. Among all types of DSLR effects, the narrow depth of field (DoF) or so called bokeh probably arouses most interest. Today's smartphones try to overcome the physical lens and sensor limitations by introducing computational methods that utilize a depth map to synthesize the narrow DoF effect from all-in-focus images. However, a high quality depth map remains to be the key differentiator between computational bokeh and DSLR optical bokeh. Empowered by a novel wavelet synthesis network architecture, we have greatly narrowed the gap between DSLR and smartphone camera in terms of the bokeh more than ever before. We describe three key Modern smartphone cameras can match traditional digital single lens reflex (DSLR) cameras in many areas thanks to the introduction of camera arrays and multi-frame processing. Among all types of DSLR effects, the narrow depth of field (DoF) or so called bokeh probably arouses most interest. Today's smartphones try to overcome the physical lens and sensor limitations by introducing computational methods that utilize a depth map to synthesize the narrow DoF effect from all-in-focus images. However, a high quality depth map remains to be the key differentiator between computational bokeh and DSLR optical bokeh. Empowered by a novel wavelet synthesis network architecture, we have narrowed the gap between DSLR and smartphone camera in terms of bokeh more than ever before. We describe three key enablers of our bokeh solution: a synthetic graphics engine to generate training data with precisely prescribed characteristics that match the real smartphone captures, a novel wavelet synthesis neural network (WSN) architecture to produce unprecedented high definition disparity map promptly on smartphones, and a new evaluation metric to quantify the quality of the disparity map for real images from the bokeh rendering perspective. Experimental results show that the disparity map produced from our neural network achieves much better accuracy than the other state-of-the-art CNN based algorithms. Combining the high resolution disparity map with our rendering algorithm, we demonstrate visually superior bokeh pictures compared with existing top rated flagship smartphones listed on the DXOMARK mobiles.

Capacitive sensing technology is ubiquitous in today's electronic devices. This paper proposes a novel architecture for the design of an ultra low power self-capacitive touch sensing analog front end (AFE) by exploiting the sparsity of simultaneous touches with respect to the number of sensor nodes. It is possible to significantly reduce the complexity and the power consumption of the AFE by migrating the computational burden to the digital processor which usually have surplus computational power. Based on the 1-bit compressive sensing theory, the ADC(s) in the AFE can be replaced by a single comparator. The number of measurements required in order to resolve the touch positions is related to the number of simultaneous touches rather than the number of sensor nodes. Detailed AFE architecture and the capacitance measurement process will be presented along with a corresponding digital reconstruction algorithm run by the digital processor.

This paper proposes a randomly jittered temporal sampling scheme with a successive approximation register (SAR) ADC. The sampling time points are jittered around a uniform sampling clock. A control logic is implemented on a traditional SAR ADC to force it to terminate the sample conversion before reaching the full precision at the jittered time points. As a result, variable word length data samples are produced by the SAR converter. Based on a discrete jittered random sampling theory, this paper analyzes the impact of the random jitters and the resulting randomized quantization noise for a class of sparse or compressible signals. A reconstruction algorithm called Successive Sine Matching Pursuit (SSMP) is proposed to recover spectrally sparse signals when sampled by the proposed SAR ADC at a sub-Nyquist rate.

This paper proposes a new perspective on the relationship between the sampling and aliasing. Unlike the uniform sampling case, where the aliases are simply periodic replicas of the original spectrum, random sampling theory shows that the randomization of sampling intervals shapes the aliases into a noise floor in the sampled spectrum. New insights into both the Fourier random sampling problem and Compressive Sensing theory can be obtained using the theoretical framework of random sampling. This paper extends the theory of continuous time random sampling to deal with random discrete intervals generated from a clock. A key result is established to relate the discrete probability distribution of the sampling intervals to the power spectrum of the aliasing noise. Based on the proposed theory, a generic discrete random sampling hardware architecture is also proposed for sampling and reconstructing a class of spectrally sparse signals at an average rate significantly below the Nyquist rate of the signal.

This paper presents a novel digital blind calibration method for time interleaved analog to digital converters (TIADCs). A simple cost function based on the cross-correlation of channel statistics is used to derive a steepest descent algorithm for the compensation of timing mismatch errors. Instead of calibrating the timing mismatches independently for each channel, only one adaptation channel needs to be calibrated within a closed loop. The calibration of the rest of the channels can be coordinated according to a scaling relationship established during an initialization stage. As a result, both the computational complexity and convergence speed of the proposed algorithm can be improved significantly with little loss in calibration performance.

This paper proposes a general structure for FIR filters with adjustable magnitude and phase responses controlled by a few parameters. The Farrow structure which uses one parameter to control the fractional delay of an FIR filter can be viewed as special case. A filter bank structure consisting of different types of linear phase differentiators forms the basis of the structure. The filter bank outputs are combined with coefficients derived from a polynomial expansion of the desired frequency response. The magnitude and phase responses are controlled by synthesizing the polynomial coefficients from the small set of control parameters. A new optimal polynomial approximation strategy is also proposed to better approximate the family of target frequency responses.

Loudspeakers in portable consumer electronic devices are frequently small in size. Due to the low sensitivity of their drive units, they are pushed to their power handling and mechanical limits by powerful amplifiers in an attempt to reach high volumes. To protect against excessive diaphragm excursions, a model based algorithm is proposed which regulates the voltage input signal to the loudspeaker while minimizing unnecessary system interventions.

Improved capacitive touch screen responsiveness can be achieved at the expense of the touch screen controller analog hardware complexity and power consumption. This paper proposes a compressive sensing based approach to exploit the sparsity of simultaneous touches (e.g., 10 or less per person) with respect to the number of sensor nodes (e.g., 100s) to achieve similar levels of responsiveness with lower levels of analog complexity and power consumption. This is done by showing that the number of measurements required for touch detection is related to the number of touches rather than the number of nodes.

This paper proposes a compressive sampling scheme based on random temporal sampling using a successive approximation register (SAR) ADC architecture. Variable wordlength data samples at random sampling times would be produced by the SAR converter, so a modified reconstruction algorithm is proposed to recover signals that are sparse or compressible in a known basis. The modified reconstruction algorithm addresses the variable wordlength resolution issue of SAR ADC samples introduced by the random sampling scheme. We demonstrate that the proposed sampling and reconstruction scheme performs significantly better compared with uniform sampling on the same SAR ADC architecture.

Digital filters with adjustable bandwidth(s) are generally desirable in many applications like audio processing and telecommunication. This paper proposes a generalized Farrow structure for adjustable bandwidth linear-phase FIR filters designed under a minimax design criterion. The bandwidth of the proposed filter structure can be continuously adjusted with an updating routine that only involves a few multiplications and additions. Moreover, the generalized structure can be designed to effectively reduce the dynamic range of basis filter coefficients, which is desirable when making a fixed-point implementation on FPGAs.

This paper deals with location estimation in the presence of unreliable measurements (outliers) in large scale geolocation systems. In order to find a way of identifying these outliers, we propose an adaptive RANSAC (Random Sample Consensus) algorithm that maintains a consensus set which iteratively includes inliers and rejects outliers. Numerical experiments are performed to apply the proposed algorithm to linearized TDOA (time difference of arrival) and AOA (angle of arrival) geolocation systems to quantify the location performance versus the least-squares estimate.