Beyond the hubble sequence – exploring galaxy morphology with unsupervised machine learning

Cheng, Ting-Yun; Huertas-Company, M.; Conselice, Christopher J.; Aragón-Salamanca, Alfonso; Robertson, Brant; Ramachandra, Nesar

doi:10.1093/mnras/stab734

Cited by 55 publications

(38 citation statements)

References 46 publications

(63 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…How-ever, currently the deep learning approach is unable to precisely classify the different types of tidal features, unless it is trained with large sample of images that have been previously precisely annotated. Unsupervised techniques could offer a solution to this problem (e.g., Martin et al 2020;Uzeirbegovic et al 2020;Spindler et al 2021;Cheng et al 2021), although there may be less control over the output.…”

Section: Introductionmentioning

confidence: 99%

Characterization of low surface brightness structures in annotated deep images

Sola

Duc

Richards

et al. 2022

A&A

View full text Add to dashboard Cite

Context. The identification and characterization of low surface brightness (LSB) stellar structures around galaxies such as tidal debris of ongoing or past collisions is essential to constrain models of galactic evolution. So far most efforts have focused on the numerical census of samples of varying sizes, either through visual inspection or more recently with deep learning. Detailed analyses including photometry have been carried out for a small number of objects, essentially because of the lack of convenient tools able to precisely characterize tidal structures around large samples of galaxies. Aims. Our goal is to characterize in detail, and in particular obtain quantitative measurements, of LSB structures identified in deep images of samples consisting of hundreds of galaxies. Methods. We developed an online annotation tool that enables contributors to delineate the shapes of diffuse extended stellar structures with precision, as well as artifacts or foreground structures. All parameters are automatically stored in a database which may be queried to retrieve quantitative measurements. We annotated LSB structures around 352 nearby massive galaxies with deep images obtained with the Canada-France-Hawaii Telescope as part of two large programs: Mass Assembly of early-Type GaLAxies with their fine Structures (MATLAS) and Ultraviolet Near Infrared Optical Northern Survey (UNIONS)/Canada-France Imaging Survey (CFIS). Each LSB structure was delineated and labeled according to its likely nature: stellar shells, streams associated with a disrupted satellite, tails that formed in major mergers, ghost reflections, or cirrus. Results. From our database containing 8441 annotations, the area, size, median surface brightness, and distance to the host of 228 structures were computed. The results confirm the fact that tidal structures defined as streams are thinner than tails, as expected by numerical simulations. In addition, tidal tails appear to exhibit a higher surface brightness than streams (by about 1 mag), which may be related to different survival times for the two types of collisional debris. We did not detect any tidal feature fainter than 27.5 mag arcsec −2 , while the nominal surface brightness limits of our surveys range between 28.3 and 29 mag arcsec −2 , a difference that needs to be taken into account when estimating the sensitivity of future surveys to identify LSB structures. Conclusions. We compiled an annotation database of observed LSB structures around nearby massive galaxies including tidal features that may be used for quantitative analysis and as a training set for machine learning algorithms.

show abstract

Section: Introductionmentioning

confidence: 99%

Characterization of low surface brightness structures in annotated deep images

Sola

Duc

Richards

et al. 2022

A&A

View full text Add to dashboard Cite

show abstract

“…Unsupervised learning has been adopted to avoid those disadvantages, but the corresponding classification accuracy is ∼ 10% worse than that of supervised manners (Cheng et al 2020(Cheng et al , 2021.…”

Section: First Authormentioning

confidence: 99%

Classifying Galaxy Morphologies with Few-Shot Learning

Zhang,

Zou,

et al. 2022

Preprint

View full text Add to dashboard Cite

The taxonomy of galaxy morphology is critical in astrophysics as the morphological properties are powerful tracers of galaxy evolution. With the upcoming Large-scale Imaging Surveys, billions of galaxy images challenge astronomers to accomplish the classification task by applying traditional methods or human inspection. Consequently, machine learning, in particular supervised deep learning, has been widely employed to classify galaxy morphologies recently due to its exceptional automation, efficiency, and accuracy. However, supervised deep learning requires extensive training sets, which causes considerable workloads; also, the results are strongly dependent on the characteristics of training sets, which leads to biased outcomes potentially. In this study, we attempt Few-shot Learning to bypass the two issues. Our research adopts the dataset from Galaxy Zoo Challenge Project on Kaggle, and we divide it into five categories according to the corresponding truth table. By classifying the above dataset utilizing few-shot learning based on Siamese Networks and supervised deep learning based on AlexNet, VGG 16, and ResNet 50 trained with different volumes of training sets separately, we find that few-shot learning achieves the highest accuracy in most cases, and the most significant improvement is 21% compared to AlexNet when the training sets contain 1000 images. In addition, to guarantee the accuracy is no less than 90%, few-shot learning needs ∼6300 images for training, while ResNet 50 requires ∼13000 images. Considering the advantages stated above, foreseeably, few-shot learning is suitable for the taxonomy of galaxy morphology and even for

show abstract

“…The fraction of CNN-classified Spirals with Sérsic index between 2.5 and 4 is similar to the fraction of Ellipticals in this magnitude and redshift range. This indicates that the class of lenticular galaxies which is not well defined in our training set and has ambiguous structure could possibly confuse our CNN classifier (Cheng et al 2021).…”

Section: Magnitude Bins: 16 ≤ I < 18mentioning

confidence: 99%

Galaxy Morphological Classification Catalogue of the Dark Energy Survey Year 3 data with Convolutional Neural Networks

Cheng,

Conselice,

Aragón-Salamanca

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We present in this paper one of the largest galaxy morphological classification catalogues to date, including over 20 million of galaxies, using the Dark Energy Survey (DES) Year 3 data based on Convolutional Neural Networks (CNN). Monochromatic i-band DES images with linear, logarithmic, and gradient scales, matched with debiased visual classifications from the Galaxy Zoo 1 (GZ1) catalogue, are used to train our CNN models. With a training set including bright galaxies (16 ≤ i < 18) at low redshift (z < 0.25), we furthermore investigate the limit of the accuracy of our predictions applied to galaxies at fainter magnitude and at higher redshifts. Our final catalogue covers magnitudes 16 ≤ i < 21, and redshifts z < 1.0, and provides predicted probabilities to two galaxy types -Ellipticals and Spirals (disk galaxies). Our CNN classifications reveal an accuracy of over 99% for bright galaxies when comparing with the GZ1 classifications (i < 18). For fainter galaxies, the visual classification carried out by three of the co-authors shows that the CNN classifier correctly categorises disky galaxies with rounder and blurred features, which humans often incorrectly visually classify as Ellipticals. As a part of the validation, we carry out one of the largest examination of non-parametric methods, including ∼100,000 galaxies with the same coverage of magnitude and redshift as the training set from our catalogue. We find that the Gini coefficient is the best single parameter discriminator between Ellipticals and Spirals for this data set.

show abstract

Beyond the hubble sequence – exploring galaxy morphology with unsupervised machine learning

Cited by 55 publications

References 46 publications

Characterization of low surface brightness structures in annotated deep images

Characterization of low surface brightness structures in annotated deep images

Classifying Galaxy Morphologies with Few-Shot Learning

Galaxy Morphological Classification Catalogue of the Dark Energy Survey Year 3 data with Convolutional Neural Networks

Contact Info

Product

Resources

About