ImageNet image recognition contest

  • 4 Replies
  • 2153 Views
*

DemonRaven

  • Trusty Member
  • ********
  • Replicant
  • *
  • 630
  • Disclaimer old brain @ work not liable for content
    • Chatbotfriends
ImageNet image recognition contest
« on: December 12, 2015, 06:01:56 am »
http://image-net.org/challenges/LSVRC/2015/

This is more in the line of neural nets and I know that some of you program them. IT is too late for this year but perhaps you can join in next year.


Quote
FAQ

1. Are challenge participants required to reveal all details of their methods?

Entires to ILSVRC2015 can be either "open" or "closed." Teams submitting "open" entries will be expected to reveal most details of their method (special exceptions may be made for pending publications). Teams may choose to submit a "closed" entry, and are then not required to provide any details beyond an abstract. The motivation for introducing this division is to allow greater participation from industrial teams that may be unable to reveal algorithmic details while also allocating more time at the ICCV15 ImageNet and MS COCO Visual Recognition Challenges Joint Workshop to teams that are able to give more detailed presentations. Participants are strongly encouraged to submit "open" entires if possible.

2. Can additional images or annotations be used in the competition?

Entires submitted to ILSVRC2015 will be divided into two tracks: "provided data" track (entries only using ILSVRC2015 images and annotations from any aforementioned tasks in both main competitions and taster competitions -- different from ILSVRC 2014), and "external data" track (entries using any outside images or annotations). Any team that is unsure which track their entry belongs to should contact the organizers ASAP. Additional clarifications will be posted here as needed.

3. Is there still image classification task?

There will be no more classification task on the ILSVRC 2012 classification/localization dataset, instead teams have to submit results in the localization format (meaning both class labels and bounding boxes). If they choose to do so, they can return the full image as their guess for the object bounding box. If teams are still only interested in image level classification problem, they can participate the scene classification challenge.

4. How many entries can each team submit per competition?

Participants who have investigated several algorithms may submit one result per algorithm (up to 5 algorithms). Changes in algorithm parameters do not constitute a different algorithm (following the procedure used in PASCAL VOC).
So sue me

*

8pla.net

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1302
  • TV News. Pub. UAL (PhD). Robitron Mod. LPC Judge.
    • 8pla.net
Re: ImageNet image recognition contest
« Reply #1 on: December 12, 2015, 01:26:08 pm »
Here is my prototype algorithm for next year to compute the error in the image:

First suggestion, an A.I.M.L compatible version...

Code
<math xmlns="http://www.w3.org/1998/Math/MathML" display="block">
  <mi>e</mi>
  <mo>=</mo>
  <mfrac>
    <mn>1</mn>
    <mi>n</mi>
  </mfrac>
  <mo>&#x22C5;<!-- â‹… --></mo>
  <munder>
    <mo>&#x2211;<!-- ∑ --></mo>
    <mi>k</mi>
  </munder>
  <mi>m</mi>
  <mi>i</mi>
  <msub>
    <mi>n</mi>
    <mrow class="MJX-TeXAtom-ORD">
      <mi>i</mi>
    </mrow>
  </msub>
  <mi>m</mi>
  <mi>i</mi>
  <msub>
    <mi>n</mi>
    <mrow class="MJX-TeXAtom-ORD">
      <mi>m</mi>
    </mrow>
  </msub>
  <mi>m</mi>
  <mi>a</mi>
  <mi>x</mi>
  <mo fence="false" stretchy="false">{</mo>
  <mi>d</mi>
  <mo stretchy="false">(</mo>
  <msub>
    <mi>c</mi>
    <mi>i</mi>
  </msub>
  <mo>,</mo>
  <msub>
    <mi>C</mi>
    <mi>k</mi>
  </msub>
  <mo stretchy="false">)</mo>
  <mo>,</mo>
  <mi>f</mi>
  <mo stretchy="false">(</mo>
  <msub>
    <mi>b</mi>
    <mi>i</mi>
  </msub>
  <mo>,</mo>
  <msub>
    <mi>B</mi>
    <mrow class="MJX-TeXAtom-ORD">
      <mi>k</mi>
      <mi>m</mi>
    </mrow>
  </msub>
  <mo stretchy="false">)</mo>
  <mo fence="false" stretchy="false">}</mo>
</math>

Lisp Version (Computer Algebra System)

First test run (see attached image for results):


Code
lsum (i/n, k, rootsof (min,min[m],max[d(c,C[k]),f(b,B[km])]));

This is an early prototype for discussion purposes.  It is not at all finalized.
Please comment.  We have a whole year to clean this up or start from scratch.

« Last Edit: December 12, 2015, 02:01:33 pm by 8pla.net »
My Very Enormous Monster Just Stopped Using Nine

*

DemonRaven

  • Trusty Member
  • ********
  • Replicant
  • *
  • 630
  • Disclaimer old brain @ work not liable for content
    • Chatbotfriends
Re: ImageNet image recognition contest
« Reply #2 on: December 12, 2015, 08:24:12 pm »
I am not really familiar with neural nets other then i know what they are and what they do. Perhaps one of the others that are  more knowledgeable can help.
So sue me

*

8pla.net

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1302
  • TV News. Pub. UAL (PhD). Robitron Mod. LPC Judge.
    • 8pla.net
Re: ImageNet image recognition contest
« Reply #3 on: December 13, 2015, 11:29:32 am »
One concern I have about the contest,
is that it may be limited to Java or .NET
virtual machines.

I'm not comfortable with some framework
capturing my programming talent before
it turns it into bytecode.

I want to program computers directly.
My Very Enormous Monster Just Stopped Using Nine

*

ivan.moony

  • Trusty Member
  • ************
  • Bishop
  • *
  • 1723
    • mind-child
Re: ImageNet image recognition contest
« Reply #4 on: December 13, 2015, 12:32:29 pm »
I had an idea about photo recognition software that uses OCR technique. When you fill up training data, you prepare it by detecting edges in training photos and reducing them to matrix bitmap of about 20 x 20 pixels. That way you get reduced outlines that you can compare to reduced outlines of photos you want to recognize. On recognition, you have to scale, squash, translate and rotate training data when comparing it to a photo, and that takes many combinations, resulting in algorithm slowdown. If the match is above 70%, you have a recognition candidate.

Doesn't sound fancy like "neural network", but it should work.

 


OpenAI Speech-to-Speech Reasoning Demo
by ivan.moony (AI News )
March 28, 2024, 01:31:53 pm
Say good-bye to GPUs...
by MikeB (AI News )
March 23, 2024, 09:23:52 am
Google Bard report
by ivan.moony (AI News )
February 14, 2024, 04:42:23 pm
Elon Musk's xAI Grok Chatbot
by MikeB (AI News )
December 11, 2023, 06:26:33 am
Nvidia Hype
by 8pla.net (AI News )
December 06, 2023, 10:04:52 pm
How will the OpenAI CEO being Fired affect ChatGPT?
by 8pla.net (AI News )
December 06, 2023, 09:54:25 pm
Independent AI sovereignties
by WriterOfMinds (AI News )
November 08, 2023, 04:51:21 am
LLaMA2 Meta's chatbot released
by 8pla.net (AI News )
October 18, 2023, 11:41:21 pm

Users Online

278 Guests, 0 Users

Most Online Today: 363. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles