Source author record

Ming-Chang Liu

Ming-Chang Liu appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

astro-ph.EP Computer Vision astro-ph.CO astro-ph.GA astro-ph.SR

Catalog footprint

What is connected

5works

5topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Presolar stardust in asteroid Ryugu

We have conducted a NanoSIMS-based search for presolar material in samples recently returned from C-type asteroid Ryugu as part of JAXA's Hayabusa2 mission. We report the detection of all major presolar grain types with O- and C-anomalous isotopic compositions typically identified in carbonaceous chondrite meteorites: 1 silicate, 1 oxide, 1 O-anomalous supernova grain of ambiguous phase, 38 SiC, and 16 carbonaceous grains. At least two of the carbonaceous grains are presolar graphites, whereas several grains with moderate C isotopic anomalies are probably organics. The presolar silicate was located in a clast with a less altered lithology than the typical extensively aqueously altered Ryugu matrix. The matrix-normalized presolar grain abundances in Ryugu are 4.8$^{+4.7}_{-2.6}$ ppm for O-anomalous grains, 25$^{+6}_{-5}$ ppm for SiC grains and 11$^{+5}_{-3}$ ppm for carbonaceous grains. Ryugu is isotopically and petrologically similar to carbonaceous Ivuna-type (CI) chondrites. To compare the in situ presolar grain abundances of Ryugu with CI chondrites, we also mapped Ivuna and Orgueil samples and found a total of SiC grains and 6 carbonaceous grains. No O-anomalous grains were detected. The matrix-normalized presolar grain abundances in the CI chondrites are similar to those in Ryugu: 23 $^{+7}_{-6}$ ppm SiC and 9.0$^{+5.3}_{-4.6}$ ppm carbonaceous grains. Thus, our results provide further evidence in support of the Ryugu-CI connection. They also reveal intriguing hints of small-scale heterogeneities in the Ryugu samples, such as locally distinct degrees of alteration that allowed the preservation of delicate presolar material.

preprint2016arXiv

Latent Model Ensemble with Auto-localization

Deep Convolutional Neural Networks (CNN) have exhibited superior performance in many visual recognition tasks including image classification, object detection, and scene label- ing, due to their large learning capacity and resistance to overfit. For the image classification task, most of the current deep CNN- based approaches take the whole size-normalized image as input and have achieved quite promising results. Compared with the previously dominating approaches based on feature extraction, pooling, and classification, the deep CNN-based approaches mainly rely on the learning capability of deep CNN to achieve superior results: the burden of minimizing intra-class variation while maximizing inter-class difference is entirely dependent on the implicit feature learning component of deep CNN; we rely upon the implicitly learned filters and pooling component to select the discriminative regions, which correspond to the activated neurons. However, if the irrelevant regions constitute a large portion of the image of interest, the classification performance of the deep CNN, which takes the whole image as input, can be heavily affected. To solve this issue, we propose a novel latent CNN framework, which treats the most discriminate region as a latent variable. We can jointly learn the global CNN with the latent CNN to avoid the aforementioned big irrelevant region issue, and our experimental results show the evident advantage of the proposed latent CNN over traditional deep CNN: latent CNN outperforms the state-of-the-art performance of deep CNN on standard benchmark datasets including the CIFAR-10, CIFAR- 100, MNIST and PASCAL VOC 2007 Classification dataset.

preprint2016arXiv

Multiple Instance Learning Convolutional Neural Networks for Object Recognition

Convolutional Neural Networks (CNN) have demon- strated its successful applications in computer vision, speech recognition, and natural language processing. For object recog- nition, CNNs might be limited by its strict label requirement and an implicit assumption that images are supposed to be target- object-dominated for optimal solutions. However, the labeling procedure, necessitating laying out the locations of target ob- jects, is very tedious, making high-quality large-scale dataset prohibitively expensive. Data augmentation schemes are widely used when deep networks suffer the insufficient training data problem. All the images produced through data augmentation share the same label, which may be problematic since not all data augmentation methods are label-preserving. In this paper, we propose a weakly supervised CNN framework named Multiple Instance Learning Convolutional Neural Networks (MILCNN) to solve this problem. We apply MILCNN framework to object recognition and report state-of-the-art performance on three benchmark datasets: CIFAR10, CIFAR100 and ILSVRC2015 classification dataset.

preprint2014arXiv

Dense molecular cloud cores as a source of micrometer-sized grains in galaxies

Coreshine in dense molecular cloud cores (dense cores) is interpreted as evidence for micrometer-sized grains (referred to as very large grains, VLGs). VLGs may have a significant influence on the total dust amount and the extinction curve. We estimate the total abundance of VLGs in the Galaxy, assuming that dense cores are the site of VLG formation. We find that the VLG abundance relative to the total dust mass is roughly $ϕ_\mathrm{VLG}\sim 0.01(1-ε)/ε(τ_\mathrm{SF}/5\times 10^9~\mathrm{yr})^{-1} (f_\mathrm{VLG}/0.5)(t_\mathrm{shat}/10^8~\mathrm{yr})$, where $ε$ is the star formation efficiency in dense cores, $τ_\mathrm{SF}$ the timescale of gas consumption by star formation, $f_\mathrm{VLG}$ the fraction of dust mass eventually coagulated into VLGs in dense cores, and $t_\mathrm{shat}$ the lifetime of VLGs (determined by shattering). Adopting their typical values for the Galaxy, we obtain $ϕ_\mathrm{VLG}\sim 0.02$--0.09. This abundance is well below the value detected in the heliosphere by Ulysses and Galileo, which means that local enhancement of VLG abundance in the solar neighborhood is required if the VLGs originate from dense cores. We also show that the effects of VLGs on the extinction curve are negligible even with the upper value of the above range, $ϕ_\mathrm{VLG}\sim 0.09$. If we adopt an extreme value, $ϕ_\mathrm{VLG}\sim 0.5$, close to that inferred from the above spacecraft data, the extinction curve is still in the range of the variation in Galactic extinction curves, but is not typical of the diffuse ISM.

preprint2011arXiv

Protosolar Irradiation in the Early Solar System: Clues from Lithium and Boron Isotopes

We report Li and B isotopic compositions of 10 Spinel-HIBonite spherules (SHIBs) separated from the Murchison meteorite, in order to understand their irradiation history in the early Solar System. The extremely low Be concentrations in SHIBs preclude detection of extinct 10Be, but instead allow for a search of the original Li and B isotopic ratios of the grains, as these isotopes are sensitive indicators for irradiation. We found that some of the SHIBs carried sub-chondritic 7Li/6Li and supra-chondritic 10B/11B ratios. Considering two possible irradiation scenarios that could have occurred in the early Solar System, irradiation of hibonite solids followed by addition of isotopically normal Li and B seems to be the most plausible explanation for the observed Li and B isotope ratios.