Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching.

Weakly supervised semantic segmentation (WSSS) is a challenging yet important research field in vision community. In WSSS, the key problem is to generate high-quality pseudo segmentation masks (PSMs). Existing approaches mainly depend on the discriminative object part to generate PSMs, which would inevitably miss object parts or involve surrounding image background, as the learning process is unaware of the full object structure. In fact, both the discriminative object part and the full object structure are critical for deriving of high-quality PSMs. To fully explore these two information cues, we build a novel end-to-end learning framework, alternate self-dual teaching (ASDT), based on a dual-teacher single-student network architecture. The information interaction among different network branches is formulated in the form of knowledge distillation (KD). Unlike the conventional KD, the knowledge of the two teacher models would inevitably be noisy under weak supervision. Inspired by the Pulse Width (PW) modulation, we introduce a PW wave-like selection signal to alleviate the influence of the imperfect knowledge from either teacher model on the KD process. Comprehensive experiments on the PASCAL VOC 2012 and COCO-Stuff 10K demonstrate the effectiveness of the proposed ASDT framework, and new state-of-the-art results are achieved.

Generalized Weakly Supervised Object Localization.

With the goal of learning to localize specific object semantics using the low-cost image-level annotation, weakly supervised object localization (WSOL) has been receiving increasing attention in recent years. Although existing literatures have studied a number of major issues in this field, one important yet challenging scenario, where the test object semantics may appear in the training phase (seen categories) or never been observed before (unseen categories), is still beyond the exploration of the existing works. We define this scenario as the generalized WSOL (GWSOL) and make a pioneering effort to study it in this article. By leveraging attribute vectors to associate seen and unseen categories, we involve threefold modeling components, i.e., the class-sensitive modeling, semantic-agnostic modeling, and content-aware modeling, into a unified end-to-end learning framework. Such design enables our model to recognize and localize unconstrained object semantics, learn compact and discriminative features that could represent the potential unseen categories, and customize content-aware attribute weights to avoid localizing on misleading attribute elements. To advance this research direction, we contribute the bounding-box manual annotations to the widely used AwA2 dataset and benchmark the GWSOL methods. Comprehensive experiments demonstrate the effectiveness of our proposed learning framework and each of the considered modeling components.

Weakly Supervised Object Detection Using Proposal- and Semantic-Level Relationships.

In recent years, weakly supervised object detection has attracted great attention in the computer vision community. Although numerous deep learning-based approaches have been proposed in the past few years, such an ill-posed problem is still challenging and the learning performance is still behind the expectation. In fact, most of the existing approaches only consider the visual appearance of each proposal region but ignore to make use of the helpful context information. To this end, this paper introduces two levels of context into the weakly supervised learning framework. The first one is the proposal-level context, i.e., the relationship of the spatially adjacent proposals. The second one is the semantic-level context, i.e., the relationship of the co-occurring object categories. Therefore, the proposed weakly supervised learning framework contains not only the cognition process on the visual appearance but also the reasoning process on the proposal- and semantic-level relationships, which leads to the novel deep multiple instance reasoning framework. Specifically, built upon a conventional CNN-based network architecture, the proposed framework is equipped with two additional graph convolutional network-based reasoning models to implement object location reasoning and multi-label reasoning within an end-to-end network training procedure. Comprehensive experiments on the widely used PASCAL VOC and MS COCO benchmarks have been implemented, which demonstrate the superior capacity of the proposed approach when compared with other state-of-the-art methods and baseline models.

Determination of 2-methoxyestradiol in serum samples and pharmaceutical preparations by silver nanoparticles-enhanced chemiluminescence.

Silver nanoparticles (AgNPs) exhibited better chemiluminescence (CL) catalysis activity and smaller nanoparticles have stronger catalysis ability in luminol-K3Fe(CN)6 system among the synthesized AgNPs of different size. 10±2 nm nanoparticles was used as catalysts to enhance the reaction sensitivity. It was found that the CL intensity of AgNPs-luminol-K3Fe(CN)6 was strongly inhibited in the presence of 2-methoxyestradiol (2-ME) and the relative CL intensity was in linear correlation with the concentration of 2-ME. Thus, the silver nanoparticles-enhanced CL method for the determination of 2-ME was developed. The proposed method has a detection limit (3 Sb/K) of 5.0×10(-10) mol L(-1) with a relative standard deviation of 0.75% for 5.0×10(-8) mol L(-1) 2-ME. The method was successfully applied for determination of 2-ME in human serum and pharmaceutical preparations. The possible CL reaction mechanism was also discussed briefly. Oxygen radicals played an important role in the catalytic process.

A novel chemiluminescence assay of mitoxantrone based on diperiodatocuprate(III) oxidation.

A novel and strong chemiluminescence (CL) of luminol with diperiodatocuprate (K5[Cu(HIO6)2]) was observed in alkaline medium. After the addition of mitoxantrone (MTX) into this system, the CL intensity could be greatly inhibited by MTX. Based on the phenomenon, a sensitive CL method was established for analysis of MTX combining with flow injection technology. Under optimum experimental conditions, the CL intensity was linearly related to the logarithm concentration of MTX from 5.0×10(-9)-1.0×10(-7) g/ml with the detection limit of 1.1×10(-9) g/ml (S/N=3). The relative standard deviation was 1.2% for 5.0×10(-8) g/ml of MTX. The proposed method was successfully applied for determination of MTX in pharmaceutical preparations and biological fluids. The possible CL reaction mechanism was also discussed briefly.

Sensitive determination of 2-methoxyestradiol in pharmaceutical preparations and serum samples using flow injection chemiluminescence.

A rapid and sensitive flow injection chemiluminescence (FI-CL) method is described for the determination of 2-methoxyestradiol (2ME) based on enhancement of the CL intensity from a potassium ferricyanide-calcein system in sodium hydroxide medium. The optimum conditions for the CL emission were investigated. Under optimized conditions, a linear calibration graph was obtained over the range 1.0 × 10(-8) to 1.0 × 10(-6) mol/L (r = 0.998) 2ME with a detection limit (3σ) of 5.4 × 10(-9) mol/L. The relative standard deviation (RSD) for 5.0 × 10(-7) mol/L 2ME was 1.7%. As a preliminary application, the proposed method was successfully applied to the determination of 2ME in injection solutions and serum samples. The possible CL mechanism was also proposed.

Synthesis and biological evaluation of a folate-targeted rhaponticin conjugate.

To improve the therapeutic effect of rhaponticin (RHA), a folate receptor (FR) targeted RHA conjugate was synthesized by utilizing a hydrophilic peptide spacer linked to folic acid (FA) via a releasable disulfide linker. This water-soluble conjugate was found to retain high affinity for FR-positive cells, and it produced specific, dose-responsive activity in vitro. Treatment of FRHA with a reducing agent indicated that the amino-reactive derivative of RHA would be released spontaneously following disulfide bond reduction within the endosomes. FRHA also proved to be active predominantly specific against FR-positive syngeneic and xenograft models in vivo, and possible curative activity resulted with minimal to moderate toxicity. The FRHA conjugate greatly enhanced the therapeutic effects and reduced the toxicity of RHA. In conclusion, FRHA represents a folate-targeted chemotherapeutic that can produce potent activity against established sc tumors. Hence, this report has a great significance in pharmacology and clinical medicine as well as methodology.