In order to deploy current computer vision (CV) models on resource-constrained low-power devices, recent works have proposed in-sensor and in-pixel computing approaches that try to partly/fully bypass the image signal processor (ISP) and yield significant bandwidth reduction between the image sensor and the CV processing unit by downsampling the activation maps in the initial convolutional neural network (CNN) layers. However, direct inference on the raw images degrades the test accuracy due to the difference in covariance of the raw images captured by the image sensors compared to the ISP-processed images used for training. Moreover, it is difficult to train deep CV models on raw images, because most (if not all) large-scale open-source datasets consist of RGB images. To mitigate this concern, we propose to invert the ISP pipeline, which can convert the RGB images of any dataset to its raw counterparts, and enable model training on raw images. We release the raw version of the COCO dataset, a large-scale benchmark for generic high-level vision tasks. For ISP-less CV systems, training on these raw images result in a 7.1% increase in test accuracy on the visual wake works (VWW) dataset compared to relying on training with traditional ISP-processed RGB datasets. To further improve the accuracy of ISP-less CV models and to increase the energy and bandwidth benefits obtained by in-sensor/in-pixel computing, we propose an energy-efficient form of analog in-pixel demosaicing that may be coupled with in-pixel CNN computations. When evaluated on raw images captured by real sensors from the PASCALRAW dataset, our approach results in a 8.1% increase in mAP. Lastly, we demonstrate a further 20.5% increase in mAP by using a novel application of few-shot learning with thirty shots each for the novel PASCALRAW dataset, constituting 3 classes.
Systematic changes, first to the structure of the catalyst scaffold and then to the ligating groups, are used to fine tune supramolecular catalysts to achieve high regioselectivity (95–98%) and high enantioselectivity (94–97% ee) across a series of meta-substituted styrenes varying in electronic character.
Soil fungi are involved in the decomposition of organic matter, alter soil structure and physicochemical properties and drive the material cycle and energy flow in terrestrial ecosystems. Grazing exclusion is one of the most common measures to restore degraded grasslands worldwide. However, changes in soil fungal community characteristics during grazing exclusion in different types of grasslands are unknown, and in particular, there is a lack of experimental evidence on the effects of grassland type and grazing exclusion on soil fungal diversity. Here, we investigate the effects of a 9-year grazing exclusion on soil properties, fungal community composition and diversity in three grassland types (Temperate desert, Temperate steppe and Mountain meadow). The results showed that soil fungal community composition differed between grassland types, and that grazing exclusion, grassland type and the interaction between the two did not significantly affect soil fungal α-diversity and composition, but significantly altered fungal β-diversity. Overall, our results highlight the importance of soil nutrient content on fungal diversity, particularly soil total nitrogen, total phosphorus and organic carbon, and provide key evidence to reveal the mechanisms by which soil fungi respond to grazing exclusion in different grassland types.
The synchronous detection and regulation of microRNAs (miRNAs) are essential for the early tumor diagnosis and treatment but remains a challenge. An integrative DNA tetrahedral nanomachine was self-assembled for sensitive detection and negative feedback regulation of intracellular miRNAs. This nanomachine comprised a DNA tetrahedron nanostructure as the framework, and a miRNA inhibitor-controlled allosteric DNAzyme as the core. The DNA tetrahedron brought the DNAzyme and the substrate in spatial proximity and facilitated the cellular uptake of DNAzyme. In allosteric regulation of DNAzyme, the locked tetrahedral DNAzyme (L-tetra-D) and active tetrahedral DNAzyme (A-Tetra-D) were controlled by miRNA inhibitor. The combination of miRNA inhibitor and target could trigger the conformational change from L-tetra-D to A-Tetra-D. A-Tetra-D cleaved the substrate and released fluorescence for intracellular miRNA biosensing. The DNA tetrahedral nanomachine showed excellent sensitivity (with detection limit down to 0.77 pM), specificity (with one-base mismatch discrimination), biocompatibility and stability. Simultaneously, miRNA stimulus-unlocked inhibitor introduced by our nanomachine exhibited the synchronous regulation of target cells, of which regulatory performance has been verified by the upregulated levels of downstream genes/proteins and the increased cellular apoptosis. Our study demonstrated that the DNA tetrahedral nanomachine is a promising biosense-and-treat tool for the synchronous detection and regulation of intracellular miRNA, and is expected to be applied in the early diagnosis and tailored management of cancers.
Background: We have already known that idiopathic granulomatous mastitis (IGM) is a rare benign chronic inflammatory disorder that can clinically mimic breast carcinoma, especially affects parous women of childbearing age, but there is little literature to report about pregnancy associated granulomatous mastitis (PAGM). The aim of our study is to report and describe the clinical signs, managements, clinical course, and clinical outcomes after treatment of PAGM in our hospital. Methods: We retrospectively analyzed 15 pregnant patients who were diagnosed as PAGM in our hospital collected from December 2018 to December 2020 by reviewing medical records and questionnaire survey, including the patients' characteristics, clinical presentations, microbiological workups, tissue pathology, treatment modalities, outcomes, and follow-up data. Results: The mean age of these patients at diagnosis was 30.5 (range 24-35) years. All patients had one birth before, and had at least two gravida times, 6 of them (40%) had three gravida times, and only one of them had four gravida times at diagnosis. The mean weeks of gestational age were 23.7 (range 4-37) weeks. Two patients' BMI were greater than 30, which were considered obese. The mean time to presentation since last delivery was 38.4 (range 19-78) months. All patients had a history of breastfeeding; the average breastfeeding time was 12.97 months. Just 2 of them were diagnosed with lactational mastitis before. One patient smoked before, 1 patient had oral contraceptive pills before, 4 patients had breast trauma recently, 5 patients had positive bacterial culture of pyogenic fluids, 3 patients had nipple retraction, 6 patients had abnormal humoral immunity, shown as elevated C3 or C4, and 2 patients had elevated serum prolactin. All patients presented as a breast mass with pain; two of them had erythema nodosum and oligoarthritis. Nearly all patients had unilateral lesion. The mean follow-up was 11 (range 1-24) months. Thirteen patients gave birth to a healthy baby, and all babies had a healthy growth and development. Almost all patients chose observation during pregnancy. Nine patients demonstrated complete remission, five of them underwent surgery after steroids and/or antibiotics, one patient had observation alone, two chose postpartum steroids alone, and the last one chose postpartum antibiotics alone. The average time to complete remission was 11.2 (range 7-18) months. Conclusions: In general, PAGM is a much rare disorder which has onset during pregnancy, and mainly happens in the second trimester and the third trimester. PAGM patients were all parous women and generally within 5 years of their last pregnancy, also with uncertain etiology and pathogenesis. Observational therapy during pregnancy for PAGM is reliable and feasible.
Oral squamous cell carcinoma represents 90% of all oral cancers. Recurrence prevention remains an important prognostic factor in patients with oral squamous cell carcinoma, and the recovery of the oral epithelium post-surgery is still a challenge. Thus, there is an urgent need to develop a smart carrier material to realize the spatiotemporally controlled release of anticancer drugs, instead of multiple oral administrations, for recurrence prevention and promoting the reconstruction of injured epithelial tissues. Here, we developed a multi-layered nanofiber patch capable of the photothermal-triggered release of low-molecular-weight fucoidan (LMWF) from the sandwiched layer, together with electrospun fibers as the backing and top layers. The sandwiched layer was made of phase-change materials loaded with indocyanine green, a photosensitive dye, for the localized release of LMWF in response to near-infrared irradiation. We showed that the on-demand release of LMWF was able to kill oral cancer cells effectively. Furthermore, adding acellular dermal matrix to the top nanofiber layer improved the proliferation of human oral keratinocytes, while the hydrophobic back layer served as a barrier to prevent loss of the drug. Taken together, this study provides a feasible and smart material system for killing oral squamous cancer cells together with the recovery of oral epithelium.
The photophysical process of localized surface plasmon resonance (LSPR) is, for the first time, exploited for broadband photon harvesting in photo-regulated controlled/living radical polymerization. Efficient macromolecular synthesis was achieved under illumination with light wavelengths extending from the visible to the near-infrared regions. Plasmonic Ag nanostructures were in situ generated on Ag3 PO4 photocatalysts in a reversible addition-fragmentation chain transfer (RAFT) system, thereby promoting polymerization of various monomers following a LSPR-mediated electron transfer mechanism. Owing to the LSPR-enhanced broadband photon harvesting, high monomer conversion (>99 %) was achieved under natural sunlight within 0.8 h. The deep penetration of NIR light enabled successful polymerization with reaction vessels screened by opaque barriers. Moreover, by trapping active oxygen species generated in the photocatalytic process, polymerization could be implemented without pre-deoxygenation.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)