Sunday, March 23, 2014

Ex-Omnivision Engineers Found Image Sensor Startup

A group of engineers from Omnivision has founded Cista System Corp., based in San Jose, CA. The founding group is Zheng (Wilson) Du (former Module Development Manager at Omnivision), Jiangtao Pang (former director of Product Design at Omnivision), Guangbin Zhang (former Analog Design Manager at Omnivision). The new company has raised $5M in round A at the end of 2013 and joined MIPI Alliance with the adopter status.

It's not clear what the new company is working on, but Guangbin Zhang's LinkedIn page says that he has "invented new sensor readout structure with 5 pending patents. More fundamental innovations are on the way."

23 comments:

  1. This should be the second one.

    ReplyDelete
    Replies
    1. http://www.corporationwiki.com/California/Palo-Alto/nueva-imaging-inc/45442281.aspx

      Delete
    2. Nueva Imaging? Is this a company that was acquired by Silicon Optronics in 2012, reportedly for a symbolic price of $12,000?

      It must must be a sort of reunion, as Silicon Optronics was founded as a JV of Omnivision and Powerchip.

      Delete
  2. I just can't see the need for another 2D Imager start-up. The supply chain is so established that an incremental improvement isn't enough to break in. It has to be something so good that it is worth the investment to change and that is a very high hurdle (see Foveon and others who didn't make it).

    ReplyDelete
    Replies
    1. I love these anonymous guys who think that the only reason for being in the image sensor business is to make cell phone camera parts. That has to be near the top of the list of boring endeavors.

      Delete
    2. I don't think it's just anonymous guys. A lot of investors tend to have that mentality, and it's not just imaging. Most everything now has to do with cellphones, despite the fact that it is half a decade too late.

      Delete
    3. I hate to tell David, but the medical, industrial and military markets are becoming very accustomed to cellphone and consumer pricing for anything imaging related. You would not believe the rants I receive when I price a sensor at "industrial" pricing for a couple of hundred units annual run rate. Part of this is "consumer imager" companies, seeing their margins collapse and their company fortunes rise and fall on single design wins/losses, look for stability in non-consumer markets, and are now creating a rat-race to the bottom in industrial. So you have Aptina and OVT with industrial marketing divisions.

      Could you create a boutique house of 20 engineers doing very specialty imaging and providing the hand holding that the big guys can't? Sure you could, and it would have flat revenue, making just enough to keep all the engineers fed. So what would be the ROI to investors? There are plenty of guys willing to do imaging for the love, but investors do it for the money.

      Delete
  3. No cellphone vendor is going to buy from a start-up, especially one with no track record. There's just far too much risk. I wonder who their target customers are.

    Where are they going to fab parts that will give them leading performance? Any grad student can make an imager. You need to make a good one at competitive cost to have a business.

    ReplyDelete
  4. I think there is plenty of room for start ups in imaging. It is true that entering cell phone market place can be daunting. Starting in a niche area makes more sense for a startup. (e.g. medical imaging). But it is also true that you could become a nice acquisition target for Google, Microsoft, Apple, Amazon, or some other giant company if you have something interesting for the mainstream market. Once acquired, these companies have no trouble getting support from any image sensor or camera module company. Or, you could have a nice image sensor lifestyle company that makes a small steady profit and gradually grow your business. So many opportunities, just one lifetime. Sigh.

    ReplyDelete
  5. Good luck ! Sincerely…

    ReplyDelete
  6. These guys think that they have a good record since they served a big company. There are plenty of small cellphone makers who don't have enough support from OV, SONY, etc. They can give financial support to these startups. Even a small cell phone maker can consume 10-20M image sensors each year, so enough for a small team since major R&D has been done before. The only issue is when the success comes, they can be sued.

    I agree fully with Eric, it's never too late to do business if you have a nice idea and good business sense.

    ReplyDelete
  7. I think there is a good chance in for startups in both mainstream and niche market imaging. The only pre-condition is to deliver a tangible advantage. Just few ideas of such kind of advantage:

    - find a way of merging global shutter with small pixels. Not sure if smartphones really need global shutter, but I can imagine a nice marketing campaign showing that it's the next big thing in imaging.
    - DSLR-sized sensor with single-shot HDR having no motion or other artifacts and no sensitivity loss.
    - Complete open source sensor design that makes sure that it does not infringe on any patent. Then such a design can be freely used (copied) by any company in the world, while the original designers would nicely profit from the technical support and debugging some strange problems (similar to GNU software companies)

    ReplyDelete
  8. I have worked in both medical and the consumer worlds (2D/3D endoscopic and cameras on mobile devices). it is tough to succeed in the consumer world for sure. Give me an example from the past 5 years if you disagree. But there are a lot of opportunities in the medical world. Besides the established pillcam, CMOS based endoscopes are positioned to replace the CCD based endoscopes with smaller size and lower power consumption. Stryker or medical companies in general have been moving really slowly in these fields. On the extreme end, there is a demand for tiny camera for endoscopes for applications such as fallopian tube inspection. Awaiba is leading that field of tiny cameras.

    ReplyDelete
  9. Array imager combined with novel wafer-level optics could be another good one.

    ReplyDelete
  10. How about improving low light image quality? Security and consumer apps would certainly benefit.

    ReplyDelete
    Replies
    1. Market leaders have QE of over 70%, low color crosstalk and read noise below 2e. It's already close to the limits.

      Delete
    2. According to the whitepaper at http://www.pixpolar.com/wp-content/uploads/2013/09/PIXPOLAR_WHITEPAPER_20130929.pdf major improvements could still be made to low light image quality.

      Delete
    3. No investor can understand what is written in this white paper. I think that for them this is a real white paper = nothing on the paper. Please make a more conceptual white paper so a normal guy can capture the fundamental advantages in your technology.

      Delete
    4. The white paper is intended for image sensor specialists only. However, the outcome of the low light image quality comparison presented in the white paper can be easily understood. For example, under certain circumstances Pixpolar's image sensor technology enables 2.3 times better SNR when compared to CMOS and CCD image sensor technologies. This means basically that in the specific image area you'll have 2.3 times better image quality with Pixpolar's technology or that you need to utilize 5.3 times longer exposure time with CMOS or CCD image sensors in order to reach the same image quality.

      Delete
  11. As far as I know, there are at least six imaging startups founded or operated by ex-ov person, including Invisage, Nueva, Cista, and other three in China. There may or may not be anything special about this one. However, fair market competition is something to be encouraged.

    ReplyDelete
  12. What are those three companies in China?

    ReplyDelete
    Replies
    1. Sorry I didn't ask. If you don't see it published somewhere else, it probably means that they prefer to keep low profile for some time.

      Delete

All comments are moderated to avoid spam.