<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <div class="moz-cite-prefix">Hi Karim,</div>
    <div class="moz-cite-prefix"><br>
    </div>
    <div class="moz-cite-prefix">Hmmm..  "quality of the image database"
      seems to suggest there is some art involved.  (I like that).</div>
    <div class="moz-cite-prefix"><br>
    </div>
    <div class="moz-cite-prefix">I'm still not clear as to what is meant
      by non-positive examples.  Does each image in the training set
      need to be marked in advance somehow as to it's relevance?   I
      don't really know how this works.</div>
    <div class="moz-cite-prefix"><br>
    </div>
    <div class="moz-cite-prefix">thanks!</div>
    <div class="moz-cite-prefix">David</div>
    <div class="moz-cite-prefix"><br>
    </div>
    <div class="moz-cite-prefix"><br>
    </div>
    <div class="moz-cite-prefix">On 11/2/20 7:02 PM, Karim Virani wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAKtnkizQ1vWaHiqOWGRujtdCjoroHBOJM4_eeOFDg-y7-Rk4pw@mail.gmail.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="ltr">My AI thinks I'm a mop. COVID hair.
        <div><br>
        </div>
        <div>Pretty much all visual object detection/classification
          neural nets require a healthy number of non-positive examples.
          This looks like a fail in exactly that - the body shape always
          being under the "ball" should have lowered the probability if
          they had bald soccer players in the general set. Though I
          guess photos of bald soccer players with extremely round heads
          heading soccer balls could get a little fuzzy...</div>
        <div><br>
        </div>
        <div>Anyhow, in our limited experience, the quality of the image
          database seems to matter as much or more than the
          classification algorithm. </div>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Mon, Nov 2, 2020 at 6:13 PM
          David Anderson via DPRGlist <<a
            href="mailto:dprglist@lists.dprg.org" moz-do-not-send="true">dprglist@lists.dprg.org</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote" style="margin:0px 0px 0px
          0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
          <div bgcolor="#FFFFFF">
            <div>As long as the players don't make the same mistake
              you'll probably be ok...<br>
            </div>
            <div><br>
            </div>
            <div>My favorite part is when the ball gets kicked straight
              toward the lineman.  The camera moves very quickly and
              smoothly to track the path but ends up locked on the
              lineman's head when a player intercepts the ball.</div>
            <div><br>
            </div>
            <div>So for those of you more familiar with neural nets, is
              there a way to train it up on what not to recognize?  
              Sort of an inverse training?  Say, if the soccer ball is
              attached to a neck and shoulders, it's probablhy not a
              soccer ball?   Can you show it a bunch of pictures of
              things that are <u>not</u> soccer balls to improve it's
              recognition?<br>
            </div>
            <div><br>
            </div>
            <div>cheers,</div>
            <div>David</div>
            <div><br>
            </div>
            <div><br>
            </div>
            <div><br>
            </div>
            <div>On 11/2/20 5:38 PM, Doug Paradis wrote:<br>
            </div>
            <blockquote type="cite">
              <div dir="ltr">Same thing happens to me when I play
                lineman.</div>
              <br>
              <div class="gmail_quote">
                <div dir="ltr" class="gmail_attr">On Mon, Nov 2, 2020 at
                  1:32 PM David Anderson via DPRGlist <<a
                    href="mailto:dprglist@lists.dprg.org"
                    target="_blank" moz-do-not-send="true">dprglist@lists.dprg.org</a>>
                  wrote:<br>
                </div>
                <blockquote class="gmail_quote" style="margin:0px 0px
                  0px 0.8ex;border-left:1px solid
                  rgb(204,204,204);padding-left:1ex">Evidently using an
                  AI controlled camera that's supposed to follow the <br>
                  ball but gets confused by the lineman's bald head.<br>
                  <br>
                  <a href="https://www.youtube.com/watch?v=9zoJP2FkpgU"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">https://www.youtube.com/watch?v=9zoJP2FkpgU</a><br>
                  <br>
                  In fairness, in that lighting, his head does look a
                  lot like the ball.<br>
                  <br>
                  cheers!<br>
                  <br>
                  dpa<br>
                  <br>
                  <br>
                  _______________________________________________<br>
                  DPRGlist mailing list<br>
                  <a href="mailto:DPRGlist@lists.dprg.org"
                    target="_blank" moz-do-not-send="true">DPRGlist@lists.dprg.org</a><br>
                  <a
                    href="http://lists.dprg.org/listinfo.cgi/dprglist-dprg.org"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://lists.dprg.org/listinfo.cgi/dprglist-dprg.org</a><br>
                </blockquote>
              </div>
            </blockquote>
            <p><br>
            </p>
          </div>
          _______________________________________________<br>
          DPRGlist mailing list<br>
          <a href="mailto:DPRGlist@lists.dprg.org" target="_blank"
            moz-do-not-send="true">DPRGlist@lists.dprg.org</a><br>
          <a href="http://lists.dprg.org/listinfo.cgi/dprglist-dprg.org"
            rel="noreferrer" target="_blank" moz-do-not-send="true">http://lists.dprg.org/listinfo.cgi/dprglist-dprg.org</a><br>
        </blockquote>
      </div>
    </blockquote>
    <p><br>
    </p>
  </body>
</html>