Coded DNN Watermark: Robustness against Pruning Models Using Constant Weight Code.

Deep Neural Network (DNN) watermarking techniques are increasingly being used to protect the intellectual property of DNN models. Basically, DNN watermarking is a technique to insert side information into the DNN model without significantly degrading the performance of its original task. A pruning attack is a threat to DNN watermarking, wherein the less important neurons in the model are pruned to make it faster and more compact. As a result, removing the watermark from the DNN model is possible. This study investigates a channel coding approach to protect DNN watermarking against pruning attacks. The channel model differs completely from conventional models involving digital images. Determining the suitable encoding methods for DNN watermarking remains an open problem. Herein, we presented a novel encoding approach using constant weight codes to protect the DNN watermarking against pruning attacks. The experimental results confirmed that the robustness against pruning attacks could be controlled by carefully setting two thresholds for binary symbols in the codeword.

Evaluation of peak skin dose during percutaneous coronary intervention procedures: relationship with fluoroscopic pulse rate and target vessel.

This study aimed to evaluate the relationship between the peak skin dose (PSD) associated with radiation skin injury and the fluoroscopic pulse rate or target vessel during percutaneous coronary intervention (PCI) procedures. We consecutively included 213 patients who underwent PCI procedures. The fluoroscopic time (FT), total number of cine frames, reference air kerma (RAK), and PSD were compared between the two types of fluoroscopic pulse rates (10 and 7.5 pulses/s) and among target vessels. The total number of X-ray tube angulations for each target vessel was also investigated. The median FT was 21.5 min in the 10 pulses/s group and 19.4 min in the 7.5 pulses/s group (p = 0.068, Wilcoxon rank sum test). The median PSD in the 10 pulses/s group was 749 mGy, which was significantly higher than that in the 7.5 pulses/s group (549 mGy) (p < 0.001). The median RAK in the right coronary artery (RCA) was equivalent to that in the left anterior descending artery. However, among the target vessels, the median PSD tended to be the highest in the RCA. There was a difference in the X-ray tube angulation used depending on the target vessel. PCI in the RCA used the left anterior oblique angle more frequently than PCI in the other vessels and tended to use only one angulation. The calculated PSD was related to the target vessel of the PCI procedure, and it was also closely related to the X-ray tube angulation.

Expression of SOLOIST/MRTFB i4, a novel neuronal isoform of the mouse serum response factor coactivator myocardin-related transcription factor-B, negatively regulates dendritic complexity in cortical neurons.

Megakaryoblastic leukemia 2 (MKL2)/myocardin-related transcription factor-B (MRTFB), a serum response factor (SRF) coactivator, is an important regulator of gene expression and neuronal morphology. Here, we show that different mouse MRTFB splice isoforms, including a novel fourth MRTFB isoform named spliced neuronal long isoform of SRF transcriptional coactivator (SOLOIST)/MRTFB isoform 4 (MRTFB i4), play distinct roles in this process. SOLOIST/MRTFB i4 has a short exon that encodes 21 amino acid residues ahead of the first RPXXXEL (RPEL) motif in MRTFB isoform 3. Quantitative PCR revealed that SOLOIST/MRTFB i4 and isoform 1 were enriched in the forebrain and neurons, and up-regulated during brain development. Conversely, isoform 3 was detected in various tissues, including both neurons and astrocytes, and was down-regulated in the developing brain. Reporter assays supported the SRF-coactivator function of SOLOIST/MRTFB i4 as well as isoform 1. Acute expression of MRTFB isoform 1, but not isoform 3 or SOLOIST/MRTFB i4, in neuronal cells within 24 hr drastically increased endogenous immediate early gene [c-fos, egr1, and activity-regulated cytoskeleton-associated protein] expression, but not endogenous actinin α1, ß-actin, gelsolin, or srf gene expression measured by qPCR. Over-expression of SOLOIST/MRTFB i4 reduced the dendritic complexity of cortical neurons, whereas over-expression of isoform 1 increased this complexity. Co-expression of isoform 1 and SOLOIST/MRTFB i4 in cortical neurons revealed that isoform 1 competitively counteracted down-regulation by SOLOIST/MRTFB i4. Our findings indicate that MRTFB isoforms have unique expression patterns and differential effects on gene expression and dendritic complexity, which contribute to shaping neuronal circuits, at least in part.

[Real-time Lens Exposure Dose Measurement Using a Scintillator with Optical Fiber Dosimeter during Cardiac Catheter Operation].

In recent years, the exposure dose of the operator's eye lens during interventional radiology operations has become a problem. We therefore evaluated the feasibility of real-time lens dose measurement using scintillator with optical fiber (SOF) dosimeter. Given that the SOF dosimeter is calibrated for direct X-rays, we performed a calibration for scattered X-rays to investigate energy dependence and the accuracy of lens dose measurements. The detection limit was calculated using the Kaiser method. The SOF dosimeter and the radiophotoluminescence glass (RPLG) dosimeter were attached to the protective glasses worn by the operator, and the lens exposure dose of the operator during cardiac catheterization was measured. In the phantom experiment, the SOF dosimeter had an error rate of 5.45% based on the measured value of the ionization chamber dosimeter. The sensitivity characteristics of the SOF dosimeter were slightly reduced on the higher side of the effective energy. The difference in sensitivity was related to variations in the additional filter and energy dependency. The sensitivity difference was 18.5% at maximum. Furthermore, when the additional dose was displayed, the influence of noise on long-term measurement was considerable. Using the Kaiser method to obtain the detection limit, the accuracy of the integrated dose had SOF dosimeter error rates of 4.3% to 15.5% with respect to the integrated value of the RPLG dosimeter when calibrated by the ionization chamber dosimeter. The use of the SOF dosimeter allowed for the real-time visualization of the exposure status of the eye lens and measurements with a relatively high accuracy.

ESTIMATION OF Hp(3) TO THE EYE LENS OF INTERVENTIONAL RADIOLOGISTS-RELATION BETWEEN THE EYE LENS DOSE AND RADIOLOGIST'S HEIGHT.

The aim of the study was to estimate occupational radiation dose to the eye lens of radiologists and the dose reduction ratio of lead glasses during interventional radiology. Three interventional radiologists monitored Hp(3) using small-type optically stimulated luminescence dosemeters attached to the left inside and outside of the lead glasses with 0.07-mmPb [Hp(3)eye]. Hp(10) and Hp(0.07) were monitored, respectively, by attaching the personal dosemeter to the lead neck collar above the lead apron. The median Hp(3)eye with lead glasses and the median dose reduction ratio of lead glasses for the three radiologists were 8.02 mSv/y and 57.7%, respectively. The median Hp(3)eye without lead glasses [Hp(3)eye-w/o] for the three radiologists was 18.6 mSv/y, but Hp(3)eye-w/o for one of the radiologists was 24.1 mSv/y. Monitoring occupational radiation dose to the eye lens is important because interventional radiologists are at risk of exceeding the new dose limit.

Comparison of increased venous contrast in ischemic stroke using phase-sensitive MR imaging with perfusion changes on flow-sensitive alternating inversion recovery at 3 Tesla.

BACKGROUND: Increased venous contrast in ischemic stroke using susceptibility-weighted imaging has been widely reported, although few reports have compared increased venous contrast areas with perfusion change areas. PURPOSE: To compare venous contrast on phase-sensitive MR images (PSI) with perfusion change on flow-sensitive alternating inversion recovery (FAIR) images, and to discuss the clinical use of PSI in ischemic stroke. MATERIAL AND METHODS: Thirty patients with clinically suspected acute infarction of the middle cerebral artery (MCA) territory within 7 days of onset were evaluated. Phase-sensitive imaging (PSI), flow-sensitive alternating inversion recovery (FAIR), diffusion-weighted imaging (DWI) and magnetic resonance angiography (MRA) were obtained using 3 Tesla scanner. Two neuroradiologists independently reviewed the MR images, as well as the PSI, DWI, and FAIR images. They were blinded to the clinical data and to each other's findings. The abnormal area of each image was ultimately identified after both neuroradiologists reached consensus. We analyzed areas of increased venous contrast on PSI, perfusion changes on FAIR images and signal changes on DWI for each case. RESULTS: Venous contrast increased on PSI and hypoperfusion was evident on FAIR images from 22 of the 30 patients (73%). The distribution of the increased venous contrast was the same as that of the hypoperfused areas on FAIR images in 16 of these 22. The extent of these lesions was larger than that of lesions visualized by on DWI in 18 of the 22 patients. Hypointense signals reflecting hemorrhage and no increased venous contrast on PSI and hyperperfusion on FAIR images were found in six of the remaining eight patients (20%). Findings on PSI were normal and hypoperfusion areas were absent on FAIR images of two patients (7%). CONCLUSION: Increased venous contrast on PSI might serve as an index of misery perfusion and provide useful information.