Rethinking Cross-Domain Pedestrian Detection: A Background-Focused Distribution Alignment Framework for Instance-Free One-Stage Detectors.

Cross-domain pedestrian detection aims to generalize pedestrian detectors from one label-rich domain to another label-scarce domain, which is crucial for various real-world applications. Most recent works focus on domain alignment to train domain-adaptive detectors either at the instance level or image level. From a practical point of view, one-stage detectors are faster. Therefore, we concentrate on designing a cross-domain algorithm for rapid one-stage detectors that lacks instance-level proposals and can only perform image-level feature alignment. However, pure image-level feature alignment causes the foreground-background misalignment issue to arise, i.e., the foreground features in the source domain image are falsely aligned with background features in the target domain image. To address this issue, we systematically analyze the importance of foreground and background in image-level cross-domain alignment, and learn that background plays a more critical role in image-level cross-domain alignment. Therefore, we focus on cross-domain background feature alignment while minimizing the influence of foreground features on the cross-domain alignment stage. This paper proposes a novel framework, namely, background-focused distribution alignment (BFDA), to train domain adaptive one-stage pedestrian detectors. Specifically, BFDA first decouples the background features from the whole image feature maps and then aligns them via a novel long-short-range discriminator. Extensive experiments demonstrate that compared to mainstream domain adaptation technologies, BFDA significantly enhances cross-domain pedestrian detection performance for either one-stage or two-stage detectors. Moreover, by employing the efficient one-stage detector (YOLOv5), BFDA can reach 217.4 FPS ( 640×480 pixels) on NVIDIA Tesla V100 (7~12 times the FPS of the existing frameworks), which is highly significant for practical applications. The code from this study will be made publicly available.

Pre-inhalation of hydrogen-rich gases protect against caerulein-induced mouse acute pancreatitis while enhance the pancreatic Hsp60 protein expression.

BACKGROUND: Acute pancreatitis (AP) lacks targeted prevention and treatment measures. Some key points in the pathogenesis of AP remain unclear, such as early activation of pancreatic enzymes. Several recent reports have shown the protective effect of hydrogen on several AP animal models, and the mechanism is related to antioxidant activity. Heat shock protein 60 (Hsp60) is known to accompany pancreatic enzymes synthesis and secretion pathway of in pancreatic acinar cells, while role of hsp60 in AP remains a topic. Aim of this study was to investigate effect of hydrogen pretreatment on AP and the mechanisms, focusing on pancreatic oxidative stress and Hsp60 expression. METHODS: 80 mice were randomly assigned into four groups: HAP group, AP group, HNS group, and NS group and each group were set 3 observation time point as 1 h, 3 h and 5 h (n = 6-8). Mouse AP model was induced by intraperitoneal injection of 50 µg/kg caerulein per hour for 6 injections both in AP and HAP groups, and mice in NS group and HNS group given normal saline (NS) injections at the same way as control respectively. Mice in HAP group and HNS group were treated with hydrogen-rich gases inhalation for 3 days before the first injection of caerulein or saline, while mice in AP group and NS group in normal air condition. Histopathology of pancreatic tissue, plasma amylase and lipase, plasma IL-1 and IL-6, pancreatic glutathione (GSH) and malondialdehyde (MDA), and Hsp60 mRNA and protein expression were investigated. Comparisons were made by one-way analysis of variance. RESULTS: The pancreatic pathological changes, plasma amylase and lipase activity, and the increase of plasma IL-1 and IL-6 levels in AP mice were significantly improved by the hydrogen-rich gases pretreatment, Meanwhile, the pancreatic GSH content increased and the pancreatic MDA content decreased. And, the hydrogen-rich gases pretreatment improved the Hsp60 protein expression in pancreatic tissues of AP mice at 1 h and 5 h. CONCLUSIONS: Pre-inhalation of hydrogen-rich gases have a good protective effect on AP mice, and the possible mechanisms of reduced oxidative stress and the early increased pancreatic Hsp60 protein deserve attention.

Mammary-tumor-educated B cells acquire LAP/TGF-ß and PD-L1 expression and suppress anti-tumor immune responses.

B lymphocytes play a role in inhibiting the immune response against certain tumors, but the underlying mechanisms are poorly understood. EMT-6 mammary tumors grow well in wild-type (WT) mice but show reduced growth in B-cell-deficient µ(-/-) BALB/c mice (BCDM). WT mice demonstrate extensive B-cell infiltration into the tumor bed, reduced CD8(+) T cell and CD49(+) NK cell infiltration, and markedly reduced cytolytic T-cell response relative to BCDM. Expression of LAP/TGF-ß1, CD80, CD86 and PD-L1 is significantly increased in tumor-infiltrating B cells (TIL-B) relative to splenic B cells. LAP/TGF-ß1 expression on TIL-B progressively increased from 5.4±1.7% on day 8 to 43.1±6.1% by day 21 post tumor implantation. Co-culture of EMT-6 tumor cells with Naive-B cells ex vivo generated B cells (EMT6-B) with a similar immunophenotype to TIL-B. Purified TIL-B, or in-vitro-generated EMT6-B suppressed CD4(+), CD8(+) and CD4(+)CD25(-) T-cell proliferation, and Th1 cytokine secretion, and also suppressed purified NK-cell proliferation in response to IL-15, compared to naive splenic B cells. Acquired B regulatory function required direct tumor cell: B-cell contact, and was partially reversed by antibody to TGF-ß or PD-L1, leading to tumor rejection in vivo B-cell acquisition of a suppressive phenotype following tumor infiltration may result in profound inhibition of T-cell anti-tumor responses.