nbswords
    • Create new note
    • Create a note from template
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Write
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights New
    • Engagement control
    • Transfer ownership
    • Delete this note
    • Save as template
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Note Insights Versions and GitHub Sync Sharing URL Create Help
Create Create new note Create a note from template
Menu
Options
Engagement control Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Write
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       Owned this note    Owned this note      
    Published Linked with GitHub
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    # Selective Search for Object Recognition Paper Note ###### tags: `paper notes` `deep learning` ## 1.Intro: * 長久以來 在做圖片辨識前都要先做劃分圖片區域(segmentation) * **Semention的目的在於將圖片中的每一個物件劃分開來,以便對各個物件做更進一步的分析** * 4張圖片的各個Region都有不同的特性 * 由於圖片本身是有層次(hierarchical)的,分析的方法當然也是,因此不存在用一個單一策略去解決所有的情況。 * Exhaustive serach雖然不會漏掉任何一個object,但它所需要的計算量太大了,因此常常會做一些限制 (使用粗略的搜尋框、固定的長寬比等等)來讓它**可達成** * **但是在大部分情況,仍然還是有太多window需要search。** --- * Segmentation使用圖片的結構來產生物件的可能位置 * Exhaustive serach則抓取了所有可能的位置 * 這篇paper的目的就在於結合這兩個方法,並且增加技巧的多樣性,以應付盡可能多的情況 --- * 使用Pascal VOC detection challenge的數據集來評估SS * 主要是根據bounding boxes這方面來評估quality --- ### Main research questions * (1) 作為SS實作segmentation時,什麼是一個好的多樣化方法呢? * (2) SS在圖片中創造一小群高品質區域的時候有多有效率? * (3) 在物件辨識上,我們可以利用SS去使用更強大的分類器和appearance models嗎? --- ## 2.Related work: ### Exhaustive Search * 在本篇論文提出之前,通常使用HOG+SVM做物件辨識 * 雖然有很多[緩解其限制的方法](https://www.robots.ox.ac.uk/~vgg/publications/2009/Lampert09a/lampert09a.pdf)提出來,但每張圖片需要掃過的window至少還是有10萬個 * **物件辨識還是一直圍繞在HOG+SVM** * 補充: [HOG:通常用於物件檢測的特徵描述符號。依賴於影像內物件本身的特性而擁有梯度或邊緣方向的分佈。在每個區塊的影像內計算梯度。區域被認為是畫素網格,其中梯度由區塊內畫素強度的變化幅度和方向構成](http://alex-phd.blogspot.com/2014/03/hog.html) ### Segmentation * 之前的方法是使用單一且強大的策略來應付各種情況 * SS則是利用多種策略來應付、分割,不僅有較低的運算需求而且分別處理不同的圖片條件會使得結果更一致、更好,還因為結果一致的關係讓它是可以被評估的。 ### Other sampling method (不太重要) * 1.隨機抽樣boxes,挑選出擁有最高物件性(objectness)的當成物件,用來減少windows的數量 * 2.使用visual words的Bags-of-words model,學習各個viusal words和物件位置的關係,以預測新圖片的物件位置 小結. -- **不同於exhaustive serach, SS產生一小群類別不同的獨立物件位置** **不同於Segmentation專注在單一個最好的演算法,SS使用多種策略去處理** **不同於在隨機抽樣的boxes上學習objectness, SS使用bottom-up的分組流程去產生更好的物件位置** --- ## 3.Selective Search **核心理念:** **Capture all scales** - Selective search 把任何大小的物件都考慮進去 **Diversification** - diverse set of strategies to deal with all cases **Fast to Compute** - The goal of selective search is to yield a set of possible object locations for use in a practical object recognition framework --- ### 3-1 Selective Search by Hierarchical Grouping ![](https://i.imgur.com/S29SGkP.jpg) **起始狀態:使用[另一篇論文的方法](http://people.cs.uchicago.edu/~pff/papers/seg-ijcv.pdf)拿到一小群的起始$R$(starting regions)** 接下來使用貪心演算法迭代進行: 1.計算所有鄰近區域之間的相似度 把所有相似度都放到集合$S$中 2.把兩個最相似的區域($ri$和$rj$)合併在一起 並把原本$ri$和$rj$的相似度從$S$中刪掉 3.計算已合併區域($rt$)和鄰近區域的相似度 把新的相似度放到$S$中並把$rt$放入$R$中 結束迭代後從所有$R$中輸出$L$ * $S$(similarity):所有的相似度集合 * $R$(Regions):可能有物件的區域集合 * $ri$ , $rj$是兩個相似區域 * $rt$是合併$ri$和$rj$之後的合併區域 * $L$:物件位置的可能結果(最終輸出) ### 3-2 Diversification Strategies 利用以下三種方法來達成SS的Diversification (1) a variety of colour spaces with different invariance properties (2) different similarity measures $sij$ (3) by varying our starting regions --- #### **(1).為了能在不同的場景和亮度條件下都能解釋圖片,把前面提到Hierarchical Grouping algorithm在相同的其他性質下用在不同的顏色空間結果出來就會是這個圖** ![](https://i.imgur.com/gyoOy8p.png) * +是不變 * -是會變 * +/-是局部變 * 分數就是有3個顏色通道中有幾個不變 例如1/3就是3個通道有1個不變 --- #### **(2)-1 把顏色做成直方圖 每個通道都做一個 而每個直方圖都做出25個區間(bins)** ![](https://i.imgur.com/Rhf2ri8.jpg) ![](https://i.imgur.com/15yhzne.jpg) * The size of a resulting region is simply the sum of its constituents: size($rt$) =size($ri$) +size($rj$). **(2)-2 測量texture similarity** 在不同區塊上使用SIFT-like演算法: * 對每個colour channel的八個方向都做高斯微分 得到梯度統計 * 對每個方向和每個colour channel都使用size=10的bin擷取出來做成直方圖 ![](https://i.imgur.com/nag4bJW.png) ![](https://i.imgur.com/YAXlT35.png) **(2)-3 小區域優先合併** ![](https://i.imgur.com/7fpaQg9.png) **(2)-4 $ri$ 和$rj$ fit得有多好?** ![](https://i.imgur.com/zvvMlJU.png) **(2)-5 把上述四種方法的結果都合起來** ![](https://i.imgur.com/ou1xXAU.png) ![](https://i.imgur.com/W3Rv7Qk.jpg) * a1,a2,a3,a4都是0或者1 (有用到這個性質或沒用到) #### **(3)改變Starting region** * Different colour space * Different threshold parameter K(影響分割後的圖片大小)[(another paper)](http://people.cs.uchicago.edu/~pff/papers/seg-ijcv.pdf) --- ## 4.Object Recognition using SelectiveSearch * Selective search + SIFT + bag-of-words + SVM ![](https://i.imgur.com/LmWbsXo.jpg) 這是作者在這篇paper中提出的SS實作Object recognition方法 * 先透過SIFT等等方法提出出圖片特徵 * 再把包含ground truth的當成Positive example 與positive example重疊20-50%的當成negative sample * 為了避免near-duplicate, negative examples在選擇過程中會剔除掉彼此重疊70%的negative sample * 開始使用SVM後 加入Hard negative example (特別強的negative example)來限制 * SS用來篩選出可能位置 * SIFT用來找相似度 * 詞袋用來辨識物件 * SVM用來分類物件 ## 5.Evaluation ![](https://i.imgur.com/jFfGOVo.png) 對於每一個固定的class c 每個真實情況G 每個計算出來的L為l 比較bounding box(L) 和真實區塊(G)的最大重疊情況來算出Overlap scores 最後把這些ABO在平均起來就變成所有class的ABO (MeanABO)了 --- ### Flat versus Hierarchy: ![](https://i.imgur.com/Lz9e4qx.png) <br></br> ### Individual diversification stategies: ![](https://i.imgur.com/YcvObno.png) ## 6.Conclusion * Use a diverse set of complementary and hierarchical grouping strategies * Result show that SS superior to “objectness” * Considerable jump in quality in term of object regions * Can be successfully used to create a good Bag-of-Words based localisation and recognition system

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password

    or

    By clicking below, you agree to our terms of service.

    Sign in via Facebook Sign in via Twitter Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully