The translational equivariant nature of CNN is a reason for its great success in the field of computer vision. However, networks do not enjoy more general equivariance properties such as rotation or scaling. This limits the generalization performance of the network. In this paper, we devise a method that provides networks with equivariance with respect to translation, rotation, and scaling simultaneously. We define a convolution-like operation and ensure equivariance based on our proposed scalable Fourier-Argand representation. The method has similar efficiency as a traditional network and hardly introduces any additional learnable parameters, since it does not face the computational issue often occurs in group-convolution operator. We verified the quality of our approach in the image classification task, demonstrating the robustness and the generalization ability to both scaled and rotated inputs.

High-speed is an essential requirement for many applications of image registration. However, existing methods are usually time-consuming due to the difficulty of the task. In this paper, different from usual feature-based ideas, we convert the matching problem into an algebraic optimization task. By solving a series of quadratic optimization equations, the underlying deformation (rotation, scaling and shift) between image pairs can be retrieved. This process is extremely fast and can be performed in real-time. Experiments show that our method can achieve good performance at a much lower computation cost. When used to initialize our earlier parametric Local All-Pass (LAP) registration algorithm, the results obtained improve significantly over the state of the art.

Signal and image representations that are steerable are essential to capture efficiently directional features. However, those that are successful at achieving directional selectivity usually use too many subbands, resulting in low computational efficiency. In this paper, we propose a two-dimensional nonlinear transform that uses only two subbands to achieve rotation invariance property, and enjoys a mirror reconstruction making it similar to a “tight frame”. The two-subband structure is merged into a unique, concise, complex-valued subband that approximates a Wirtinger gradient which is naturally steerable. Complete steerability, though, is achieved by utilizing the Fourier-Argand representation, which provides a steerable filter able to estimate the amplitude and direction of image features, even in the presence of very high noise. We demonstrate the efficiency of the representation by comparing how it performs in wavelet-based denoising algorithms.

Convolutional neural networks (CNNs) have demonstrated a strong ability to extract semantics from images during object detection; however, the extracted semantics are typically have strong scale priors for a specific circumstance. In this paper, we investigate the influence of head scale and contextual information, and then propose a scale-invariant method for head detection. Our method can dynamically detect heads depending on the complexity of the image. It uses an extra feature map to represent the scale information of the spatial relationship, and then uses this feature map for auxiliary detection. Particularly, we exploit several new techniques, including contextual information, scale invariance, and hard example mining. We evaluated our method on three head datasets and achieved state-of-the-art results for the Brainwash dataset, HollywoodHeads dataset, and SCUT-HEAD dataset

This paper presents a method that can accurately detect heads especially small heads under the indoor scene. To achieve this, we propose a novel method, Feature Refine Net (FRN), and a cascaded multi-scale architecture. FRN exploits the multi-scale hierarchical features created by deep convolu- tional neural networks. The proposed channel weighting method enables FRN to make use of features alternatively and effectively. To improve the performance of small head detection, we propose a cascaded multi-scale architecture which has two detectors. One called global detector is responsible for detecting large objects and acquiring the global distribution information. The other called local detector is designed for small objects detection and makes use of the information provided by global detector. Due to the lack of head detection datasets, we have collected and labeled a new large dataset named SCUT-HEAD which includes 4405 images with 111251 heads annotated. Experiments show that our method has achieved state-of-the-art performance on SCUT-HEAD.

Most rod-shaped bacteria elongate by inserting new cell wall material into the inner surface of the cell sidewall. This is primarily performed by a highly conserved protein complex, the elongasome, which moves processively around the cell circumference and inserts long glycan strands that act as barrel-hoop-like reinforcing structures, thereby giving rise to a rod-shaped cell. However, it remains unclear how elongasome synthesis dynamics and termination events are regulated to determine the length of these critical cell-reinforcing structures. To address this, we developed a method to track individual elongasome complexes around the entire circumference of Bacillus subtilis cells for minutes-long periods using single molecule fluorescence microscopy. We found that the B. subtilis elongasome is highly processive and that processive synthesis events are frequent terminated by rapid reversal or extended pauses. We found that cellular levels of RodA regulate elongasome processivity, reversal and pausing. Our single molecule data, together with stochastic simulations, show that elongasome dynamics and processivity are regulated by molecular motor tug-of-war competition between several, likely two, oppositely oriented peptidoglycan synthesis complexes bound to the MreB filament. Our data, thus, demonstrate that molecular motor tug-of-war is a key regulator of elongasome dynamics in B. subtilis, which likely also regulates the cell shape via modulation of elongasome processivity.