|
== Uses ==
[[File:Track_Results.webm|thumb|An example of object tracking]]
Small object detection has applications in various fields such as Video [[surveillance]] (Traffic video Surveillance,<ref>{{Cite journal |lastlast1=Saran K B |last2=Sreelekha G |title=Traffic video surveillance: Vehicle detection and classification |url=httphttps://ieeexplore.ieee.org/document/7432948/ |journal=2015 International Conference on Control Communication & Computing India (ICCC) |year=2015 |___location=Trivandrum, Kerala, India |publisher=IEEE |pages=516–521 |doi=10.1109/ICCC.2015.7432948 |isbn=978-1-4673-7349-4|s2cid=14779393 }}</ref><ref>{{Cite journal |last=Nemade |first=Bhushan |date=2016-01-01 |title=Automatic Traffic Surveillance Using Video Tracking |url=https://www.sciencedirect.com/science/article/pii/S1877050916001836 |journal=Procedia Computer Science |series=Proceedings of International Conference on Communication, Computing and Virtualization (ICCCV) 2016 |language=en |volume=79 |pages=402–409 |doi=10.1016/j.procs.2016.03.052 |issn=1877-0509}}</ref> [[Content-based image retrieval|Small object retrieval]],<ref>{{Cite journal |lastlast1=Guo |firstfirst1=Haiyun |last2=Wang |first2=Jinqiao |last3=Xu |first3=Min |last4=Zha |first4=Zheng-Jun |last5=Lu |first5=Hanqing |date=2015-10-13 |title=Learning Multi-view Deep Features for Small Object Retrieval in Surveillance Scenarios |url=https://doi.org/10.1145/2733373.2806349 |journal=Proceedings of the 23rd ACM internationalInternational conferenceConference on Multimedia |series=MM '15 |___location=New York, NY, USA |publisher=Association for Computing Machinery |pages=859–862 |doi=10.1145/2733373.2806349 |isbn=978-1-4503-3459-4|s2cid=9041849 }}</ref><ref>{{Cite journal |lastlast1=Galiyawala |firstfirst1=Hiren |last2=Raval |first2=Mehul S. |last3=Patel |first3=Meet |date=2022-05-20 |title=Person retrieval in surveillance videos using attribute recognition |url=https://doi.org/10.1007/s12652-022-03891-0 |journal=Journal of Ambient Intelligence and Humanized Computing |language=en |doi=10.1007/s12652-022-03891-0 |s2cid=248951090 |issn=1868-5145}}</ref> [[Anomaly detection]],<ref>{{Cite journal |lastlast1=Ingle |firstfirst1=Palash Yuvraj |last2=Kim |first2=Young-Gab |date=2022-05-19 |title=Real-Time Abnormal Object Detection for Video Surveillance in Smart Cities |url=https://www.mdpi.com/1424-8220/22/10/3862 |journal=Sensors |language=en |volume=22 |issue=10 |pages=3862 |doi=10.3390/s22103862 |issn=1424-8220 |pmc=9143895 |pmid=35632270|bibcode=2022Senso..22.3862I |doi-access=free }}</ref> [[Maritime surveillance]], [[Aerial survey|Drone surveying]], [[Traffic flow|Traffic flow analysis]],<ref>{{Cite journal |lastlast1=Tsuboi |firstfirst1=Tsutomu |last2=Yoshikawa |first2=Noriaki |date=2020-03-01 |title=Traffic flow analysis in Ahmedabad (India) |url=https://www.sciencedirect.com/science/article/pii/S2213624X18301974 |journal=Case Studies on Transport Policy |language=en |volume=8 |issue=1 |pages=215–228 |doi=10.1016/j.cstp.2019.06.001 |s2cid=195543435 |issn=2213-624X}}</ref> and [[Video tracking|Object tracking]].
== Problems with small objects ==
* Modern-day object detection algorithms such as You Only Look Once(YOLO)<ref>{{cite arxivarXiv |lastlast1=Redmon |firstfirst1=Joseph |last2=Divvala |first2=Santosh |last3=Girshick |first3=Ross |last4=Farhadi |first4=Ali |date=2016-05-09 |title=You Only Look Once: Unified, Real-Time Object Detection |arxiveprint=1506.02640}}</ref><ref>{{cite arxivarXiv |lastlast1=Redmon |firstfirst1=Joseph |last2=Farhadi |first2=Ali |date=2016-12-25 |title=YOLO9000: Better, Faster, Stronger |eprint=arxiv.1612.08242}}</ref><ref>{{cite arxivarXiv |lastlast1=Redmon |firstfirst1=Joseph |last2=Farhadi |first2=Ali |date=2018-04-08 |title=YOLOv3: An Incremental Improvement |eprint=arxiv.1804.02767}}</ref><ref>{{cite arxivarXiv |lastlast1=Bochkovskiy |firstfirst1=Alexey |last2=Wang |first2=Chien-Yao |last3=Liao |first3=Hong-Yuan Mark |date=2020-04-22 |title=YOLOv4: Optimal Speed and Accuracy of Object Detection |eprint=arxiv.2004.10934}}</ref><ref>{{cite arxivarXiv |lastlast1=Wang |firstfirst1=Chien-Yao |last2=Bochkovskiy |first2=Alexey |last3=Liao |first3=Hong-Yuan Mark |date=2021-02-21 |title=Scaled-YOLOv4: Scaling Cross Stage Partial Network |eprint=arxiv.2011.08036}}</ref><ref>{{cite arxivarXiv |lastlast1=Li |firstfirst1=Chuyi |last2=Li |first2=Lulu |last3=Jiang |first3=Hongliang |last4=Weng |first4=Kaiheng |last5=Geng |first5=Yifei |last6=Li |first6=Liang |last7=Ke |first7=Zaidan |last8=Li |first8=Qingyuan |last9=Cheng |first9=Meng |last10=Nie |first10=Weiqiang |last11=Li |first11=Yiduo |last12=Zhang |first12=Bo |last13=Liang |first13=Yufei |last14=Zhou |first14=Linyuan |last15=Xu |first15=Xiaoming |date=2022-09-07 |title=YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications |eprint=arxiv.2209.02976}}</ref><ref>{{cite arxivarXiv |lastlast1=Wang |firstfirst1=Chien-Yao |last2=Bochkovskiy |first2=Alexey |last3=Liao |first3=Hong-Yuan Mark |date=2022-07-06 |title=YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors |eprint=arxiv.2207.02696}}</ref> heavily uses convolution layers to learn [[Feature (computer vision)|features]]. As an object passes through convolution layers, its size gets reduced. Therefore, the small object disappears after several layers and becomes undetectable.
* Sometimes, the shadow of an object is detected as a part of object itself.<ref>{{Cite journal |lastlast1=Zhang |firstfirst1=Mingrui |last2=Zhao |first2=Wenbing |last3=Li |first3=Xiying |last4=Wang |first4=Dan |date=2020-12-11 |title=Shadow Detection Of Moving Objects In Traffic Monitoring Video |url=https://ieeexplore.ieee.org/document/9338958/ |journal=2020 IEEE 9th Joint International Information Technology and Artificial Intelligence Conference (ITAIC) |volume=9 |___location=Chongqing, China |publisher=IEEE |pages=1983–1987 |doi=10.1109/ITAIC49862.2020.9338958 |isbn=978-1-7281-5244-8|s2cid=231824327 }}</ref> So, the placement of the bounding box tends to centre around a shadow rather than an object. In the case of vehicle detection, [[pedestrian]] and two-wheeler detection suffer because of this.
* At present, [[Unmanned aerial vehicle|drones]] are very widely used in aerial imagery.<ref>{{Cite journal |title=Interactive workshop "How drones are changing the world we live in" |url=httphttps://ieeexplore.ieee.org/document/7486437/ |journal=2016 Integrated Communications Navigation and Surveillance (ICNS) |year=2016 |___location=Herndon, VA |publisher=IEEE |pages=1–17 |doi=10.1109/ICNSURV.2016.7486437 |isbn=978-1-5090-2149-9|s2cid=21388151 }}</ref> They are equipped with hardware ([[sensor]]s) and software ([[algorithm]]s) that help maintain a particular stable position during their flight. In windy conditions, the drone automatically makes fine moves to maintain its position and that changes the view near the boundary. It may be possible that some new objects appear near the image boundary. Overall, these affect classification, detection, and eventually tracking accuracy.
[[File:Disp_shadow.jpg|thumb|Shadow and drone movement effect|alt=Here, both images are from same video. See, How the shadow of objects affecting detection accuracy. Also, drone's self-movement changes the scene near boundary(Refer to object "car" at bottom-left corner).]]
== Methods ==
Various methods<ref>{{Cite webjournal |title=An Evaluation of Deep Learning Methods for Small Object Detection |urljournal=https://www.hindawi.com/journals/jece/2020/3189691/Journal |access-date=2022-09-14of Electrical and Computer Engineering |websiteyear=www.hindawi.com2020 |language=en |doi=10.1155/2020/3189691|doi-access=free |last1=Nguyen |first1=Nhat-Duy |last2=Do |first2=Tien |last3=Ngo |first3=Thanh Duc |last4=Le |first4=Duy-Dinh |volume=2020 |pages=1–18 }}</ref> are available to detect small objects, which fall under three categories:
[[File:Yolov5.jpg|thumb|YOLOv5 detection result]]
[[File:Y5_sahi.jpg|thumb|YOLOv5 and SAHI interface]]
==== Choosing a data set that has small objects ====
The [[machine learning]] model's output depends on "How well it is trained."<ref name=":0">{{Cite journal |lastlast1=Gong |firstfirst1=Zhiqiang |last2=Zhong |first2=Ping |last3=Hu |first3=Weidong |date=2019 |title=Diversity in Machine Learning |url=https://ieeexplore.ieee.org/document/8717641/ |journal=IEEE Access |volume=7 |pages=64323–64350 |doi=10.1109/ACCESS.2019.2917620 |s2cid=206491718 |issn=2169-3536}}</ref> So, the data set must include small objects to detect such objects. Also, modern-day detectors, such as YOLO, rely on anchors.<ref>{{Cite web |last=Christiansen |first=Anders |date=2022-06-10 |title=Anchor Boxes — The key to quality object detection |url=https://towardsdatascience.com/anchor-boxes-the-key-to-quality-object-detection-ddf9d612d4f9 |access-date=2022-09-14 |website=Medium |language=en}}</ref> Latest versions of YOLO (starting from YOLOv5<ref>{{Citation |lastlast1=Jocher |firstfirst1=Glenn |title=ultralytics/yolov5: v6.2 - YOLOv5 Classification Models, Apple M1, Reproducibility, ClearML and Deci.ai integrations |date=2022-08-17 |url=https://zenodo.org/record/7002879 |publisher=Zenodo |doi=10.5281/zenodo.3908559 |access-date=2022-09-14 |last2=Chaurasia |first2=Ayush |last3=Stoken |first3=Alex |last4=Borovec |first4=Jirka |last5=NanoCode012 |last6=Kwon |first6=Yonghye |last7=TaoXie |last8=Michael |first8=Kalen |last9=Fang |first9=Jiacong}}</ref>) uses an auto-anchor algorithm to find good anchors based on the nature of object sizes in the data set. Therefore, it is mandatory to have smaller objects in the data set.
==== Generating more data via augmentation, if required ====
==== Auto learning anchors ====
Selecting anchor size plays a vital role in small object detection.<ref>{{cite arxivarXiv |lastlast1=Zhong |firstfirst1=Yuanyi |last2=Wang |first2=Jianfeng |last3=Peng |first3=Jian |last4=Zhang |first4=Lei |date=2020-01-26 |title=Anchor Box Optimization for Object Detection |eprint=arxiv.1812.00469}}</ref> Instead of hand picking it, use algorithms that identify it based on the data set. YOLOv5 uses a [[K-means clustering|K-means algorithm]] to define anchor size.
==== Tiling approach during training and inference ====
State-of-the-art object detectors allow only the fixed size of image and change the input image size according to it. This change may deform the small objects in the image. The tiling approach<ref>{{Cite journal |lastlast1=Unel |firstfirst1=F. Ozge |last2=Ozkalayci |first2=Burak O. |last3=Cigla |first3=Cevahir |title=The Power of Tiling for Small Object Detection |url=https://ieeexplore.ieee.org/document/9025422/ |journal=2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) |year=2019 |___location=Long Beach, CA, USA |publisher=IEEE |pages=582–591 |doi=10.1109/CVPRW.2019.00084 |isbn=978-1-7281-2506-0|s2cid=198903617 }}</ref> helps when an image has a high resolution than the model's fixed input size; instead of scaling it down, the image is broken down into tiles and then used in training. The same approach is used during inference as well.
==== Feature Pyramid Network (FPN) ====
Use a feature [[Pyramid (image processing)|pyramid]] network<ref>{{cite arxivarXiv |lastlast1=Lin |firstfirst1=Tsung-Yi |last2=Dollár |first2=Piotr |last3=Girshick |first3=Ross |last4=He |first4=Kaiming |last5=Hariharan |first5=Bharath |last6=Belongie |first6=Serge |date=2017-04-19 |title=Feature Pyramid Networks for Object Detection |eprint=arxiv.1612.03144}}</ref> to learn features at a multi-scale: e.g., Twin Feature Pyramid Networks (TFPN),<ref>{{Cite journal |lastlast1=Liang |firstfirst1=Yi |last2=Changjian |first2=Wang |last3=Fangzhao |first3=Li |last4=Yuxing |first4=Peng |last5=Qin |first5=Lv |last6=Yuan |first6=Yuan |last7=Zhen |first7=Huang |title=TFPN: Twin Feature Pyramid Networks for Object Detection |url=https://ieeexplore.ieee.org/document/8995365/ |journal=2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI) |year=2019 |___location=Portland, OR, USA |publisher=IEEE |pages=1702–1707 |doi=10.1109/ICTAI.2019.00251 |isbn=978-1-7281-3798-8|s2cid=211211764 }}</ref> Extended Feature Pyramid Network (EFPN).<ref>{{cite arxivarXiv |lastlast1=Deng |firstfirst1=Chunfang |last2=Wang |first2=Mengmeng |last3=Liu |first3=Liang |last4=Liu |first4=Yong |date=2020-04-09 |title=Extended Feature Pyramid Network for Small Object Detection |eprint=arxiv.2003.07021}}</ref> FPN helps to sustain features of small objects against convolution layers.
=== Add-on techniques ===
Instead of modifying existing methods, some add-on techniques are there, which can be directly placed on top of existing approaches to detect smaller objects. One such technique is Slicing Aided Hyper Inference(SAHI).<ref>{{cite arxivarXiv |lastlast1=Akyon |firstfirst1=Fatih Cagatay |last2=Altinuc |first2=Sinan Onur |last3=Temizel |first3=Alptekin |date=2022-07-12 |title=Slicing Aided Hyper Inference and Fine-tuning for Small Object Detection |eprint=arxiv.2202.06934}}</ref> The image is sliced into different-sized multiple overlapping patches. [[Hyperparameter (machine learning)|Hyper-parameters]] define their dimensions. Then patches are resized, while maintaining the aspect ratio during fine-tuning. These patches are then provided for training the model.
=== Well-Optimised techniques for small object detection ===
Various deep learning techniques are available that focus on such object detection problems: e.g., Feature-Fused SSD,<ref>{{Cite journal |lastlast1=Cao |firstfirst1=Guimei |last2=Xie |first2=Xuemei |last3=Yang |first3=Wenzhe |last4=Liao |first4=Quan |last5=Shi |first5=Guangming |last6=Wu |first6=Jinjian |editor-first1=Junyu |editor-first2=Hui |editor-last1=Dong |editor-last2=Yu |date=2018-04-10 |title=Feature-fused SSD: fast detection for small objects |url=https://www.spiedigitallibrary.org/conference-proceedings-of-spie/10615/106151E/Feature-fused-SSD-fast-detection-for-small-objects/10.1117/12.2304811.full |journal=Ninth International Conference on Graphic and Image Processing (ICGIP 2017) |publisher=SPIE |volume=10615 |pages=381–388 |doi=10.1117/12.2304811|arxiv=1709.05054 |bibcode=2018SPIE10615E..1EC |isbn=9781510617414 |s2cid=20592770 }}</ref> YOLO-Z.<ref>{{cite arxivarXiv |lastlast1=Benjumea |firstfirst1=Aduen |last2=Teeti |first2=Izzeddin |last3=Cuzzolin |first3=Fabio |last4=Bradley |first4=Andrew |date=2021-12-23 |title=YOLO-Z: Improving small object detection in YOLOv5 for autonomous vehicles |eprint=arxiv.2112.11798}}</ref> Such methods work on "How to sustain features of small objects while they pass through convolution networks."
== Other applications ==
* Crowd counting<ref>{{Cite journal |lastlast1=Rajendran |firstfirst1=Logesh |last2=Shyam Shankaran |first2=R |title=Bigdata Enabled Realtime Crowd Surveillance Using Artificial Intelligence And Deep Learning |url=https://ieeexplore.ieee.org/document/9373133/ |journal=2021 IEEE International Conference on Big Data and Smart Computing (BigComp) |year=2021 |___location=Jeju Island, Korea (South) |publisher=IEEE |pages=129–132 |doi=10.1109/BigComp51126.2021.00032 |isbn=978-1-7281-8924-6|s2cid=232236614 }}</ref><ref>{{Cite journal |lastlast1=Sivachandiran |firstfirst1=S. |last2=Mohan |first2=K. Jagan |last3=Nazer |first3=G. Mohammed |date=2022-03-29 |title=Deep Transfer Learning Enabled High-Density Crowd Detection and Classification using Aerial Images |url=https://ieeexplore.ieee.org/document/9753982/ |journal=2022 6th International Conference on Computing Methodologies and Communication (ICCMC) |___location=Erode, India |publisher=IEEE |pages=1313–1317 |doi=10.1109/ICCMC53470.2022.9753982 |isbn=978-1-6654-1028-1|s2cid=248131806 }}</ref><ref>{{Cite journal |lastlast1=Santhini |firstfirst1=C. |last2=Gomathi |first2=V. |title=Crowd Scene Analysis Using Deep Learning Network |url=https://ieeexplore.ieee.org/document/8550851/ |journal=2018 International Conference on Current Trends towardsTowards Converging Technologies (ICCTCT) |year=2018 |pages=1–5 |doi=10.1109/ICCTCT.2018.8550851|isbn=978-1-5386-3702-9 |s2cid=54438440 }}</ref><ref>{{Cite journal |lastlast1=Sharath |firstfirst1=S.V. |last2=Biradar |first2=Vidyadevi |last3=Prajwal |first3=M.S. |last4=Ashwini |first4=B. |date=2021-11-19 |title=Crowd Counting in High Dense Images using Deep Convolutional Neural Network |url=https://ieeexplore.ieee.org/document/9663716/ |journal=2021 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics (DISCOVER) |___location=Nitte, India |publisher=IEEE |pages=30–34 |doi=10.1109/DISCOVER52564.2021.9663716 |isbn=978-1-6654-1244-5|s2cid=245707782 }}</ref>
* Vehicle re-identification<ref>{{Cite journal |lastlast1=Wang |firstfirst1=Hongbo |last2=Hou |first2=Jiaying |last3=Chen |first3=Na |date=2019 |title=A Survey of Vehicle Re-Identification Based on Deep Learning |url=https://ieeexplore.ieee.org/document/8915694/ |journal=IEEE Access |volume=7 |pages=172443–172469 |doi=10.1109/ACCESS.2019.2956172 |s2cid=209319743 |issn=2169-3536}}</ref>
* Animal detection<ref>{{Cite journal |lastlast1=Santhanam |firstfirst1=Sanjay |last2=B |first2=Sudhir Sidhaarthan |last3=Panigrahi |first3=Sai Sudha |last4=Kashyap |first4=Suryakant Kumar |last5=Duriseti |first5=Bhargav Krishna |date=2021-11-26 |title=Animal Detection for Road safety using Deep Learning |url=https://ieeexplore.ieee.org/document/9697287/ |journal=2021 International Conference on Computational Intelligence and Computing Applications (ICCICA) |___location=Nagpur, India |publisher=IEEE |pages=1–5 |doi=10.1109/ICCICA52458.2021.9697287 |isbn=978-1-6654-2040-2|s2cid=246663727 }}</ref><ref>{{Cite journal |lastlast1=Li |firstfirst1=Nopparut |last2=Kusakunniran |first2=Worapan |last3=Hotta |first3=Seiji |title=Detection of Animal Behind Cages Using Convolutional Neural Network |url=https://ieeexplore.ieee.org/document/9158137/ |journal=2020 17th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON) |year=2020 |___location=Phuket, Thailand |publisher=IEEE |pages=242–245 |doi=10.1109/ECTI-CON49241.2020.9158137 |isbn=978-1-7281-6486-1|s2cid=221086279 }}</ref><ref>{{Cite journal |lastlast1=Oishi |firstfirst1=Yu |last2=Matsunaga |first2=Tsuneo |title=Automatic detection of moving wild animals in airborne remote sensing images |url=https://ieeexplore.ieee.org/document/5654227/ |journal=2010 IEEE International Geoscience and Remote Sensing Symposium |year=2010 |pages=517–519 |doi=10.1109/IGARSS.2010.5654227|isbn=978-1-4244-9565-8 |s2cid=16812504 }}</ref><ref>{{Cite journal |lastlast1=Ramanan |firstfirst1=D. |last2=Forsyth |first2=D.A. |last3=Barnard |first3=K. |title=Building models of animals from video |url=httphttps://ieeexplore.ieee.org/document/1642665/ |journal=IEEE Transactions on Pattern Analysis and Machine Intelligence |year=2006 |volume=28 |issue=8 |pages=1319–1334 |doi=10.1109/TPAMI.2006.155 |pmid=16886866 |s2cid=1699015 |issn=0162-8828}}</ref>
* Fish detection<ref>{{Cite webjournal |title=Fish Detection Using Deep Learning |urljournal=https://www.hindawi.com/journals/acisc/2020/3738108/Applied |access-date=2022-09-14Computational Intelligence and Soft Computing |websiteyear=www.hindawi.com2020 |language=en |doi=10.1155/2020/3738108|doi-access=free |last1=Cui |first1=Suxia |last2=Zhou |first2=Yu |last3=Wang |first3=Yonghui |last4=Zhai |first4=Lujun |volume=2020 |pages=1–13 }}</ref>
== See also ==
| 