Tuesday 30 January, 2018. 17:32
I’m writing on the BART train from Millbrae back to San Francisco. I had a better sleep last night, but woke up before the alarm went off at 07:00. We got up and I had breakfast quickly and dashed out so I could catch the same train I got yesterday. There were a few more people on it today, but there were still plenty of seats.
After walking to the conference hotel, I checked in at the AV prep room, where they test to make sure your presentation laptop works with their projector system. Everything was fine, and I also asked for any tips in chairing a session, since I had to do that job for the first time this afternoon. The guy said there would be AV people running around the conference rooms to help if anyone had any problems.
Then it was off to the first session of the day. Today I spent the whole day in the Photography, Mobile, and Immersive Imaging conference, since I had my own talk to give before lunch and then a session to chair in the afternoon. First up Henry Dietz gave a talk about the development of multicamera systems at the University of Kentucky, showing off several systems they’ve developed over the years and presenting lessons learnt on how to design and engineer systems made of multiple cameras. The most crucial thing he said is to recognise that cameras these days are actually computers with sensors, not boxes for exposing film, and we need to start taking more advantage of the programmability in order to unlock their potential. I always like Henry’s talks as they are casual and fun and show off really cool stuff that his research group are doing by hacking cameras.
Next was a student talking about various camera photometric calibrations required by the Facebook 360 composite camera, to get a stitched image that matches in exposure level, black point, white balance, and colour response. Then Kevin Matherson gave an invited talk on the Microsoft HoloLens, which was part marketing spiel and then some details of the technical challenges involved in designing, building, and calibrating the system to work well enough that it wouldn’t feel clunky or induce motion sickness due to poor latency.
After the coffee break there was a very cool talk about artificially rendered changes in focal length and field of view for images. This is achieved by taking a series of photos walking towards a subject, then processing them in such a way that the field of view can vary with depth. This produces images that optical cameras can’t capture, and that combine objects at different depths in customisable relative sizes, but which still look very natural for the most part. I’d seen a paper on this by the same group a while back and thought it looked interesting, so it was good to see a live presentation about it. Then it was time for my talk. I’d prepared a lot of slides, knowing I probably wouldn’t get through them all, but I managed to complete all of the most important stuff that I wanted to get through. Although I didn’t get time for audience questions, which was a shame as I felt well prepared for anything they could throw at me.
At lunchtime was the committee meeting for the PMII conference, at Grand Harbor restaurant, just across the road. This was the same place we had the committee lunch last year, where we ended up with ridiculous amounts of food, which was still arriving as we were packing to leave and return to the conference. And yes, exactly the same thing happened this year. The new wrinkle was that every second item that came out was squid in some form or other, although the very first thing to arrive was bowls of tripe. All the best stuff – the dumplings and spring rolls and so on – came towards the end. Oh, we also discussed organisation for next year’s conference.
Written next day
Back at the talks after lunch, first up was today’s plenary, which was about scanning the interiors of buildings in 3D, using person-held or backpack-mounted scanners. This is intended to solve the problem of navigating interiors where GPS is blocked. Then the afternoon session of PMII was the session which I was chairing. I was supposed to be sharing the chair duties with Nitin Sampat, which was comforting because it was the first time I’d volunteered to do this job at a conference. But he approached me during the plenary and said I’d be fine to do it by myself!
As it turned out it was fine. All four speakers were there and contacted me before the session started so I knew what laptop connectors they needed and how to pronounce their names and so on. Two hadn’t supplied bios for the introduction on my printed sheet, so I scribbled down quick bios after asking them. The main issue was if they finished with time for questions and the audience had no questions, then I had to pose one myself. So I had to be sure to pay attention and come up with a sensible question for every talk. As it turned out, this was required after all four of the talks!
The general theme of the talks was 3D and lightfield imaging. The most interesting was by John Sasinowski of Light, talking about image quality challenges with their L16 camera that uses 16 separate camera modules in a phone-sized device to produce huge high resolution images, up to 52 MP. The engineering challenges were considerable in aligning, manufacturing, and calibrating the camera assembly. But there are also educational challenges because the quality trade offs are different to those in traditional single lens photography, and users are not used to them yet, so tend to perceive them as flaws rather than understood trade offs that they have control over.
With my session ended on time, I hustled to get back to Millbrae station in time for the 17:31 train back to San Francisco, getting me back to the hotel a bit after 18:00. M. had spent the day in the SFMOMA museum of modern art, which she had was huge, over seven floors of large galleries. It had only been three floors when I’d gone before the renovation.
We had plans to meet Mary and Casey for dinner at Oasis Grill on Market Street. But then I got an email from Mary saying she hadn’t realised it was a take-away place rather than a restaurant and suggesting a list of other possible places. I replied saying we should meet there and then decide if we want to go somewhere else, as we were heading out for some drinks and wouldn’t have data roaming capability to change plans again.
We went to the Hotel Zetta next door, which had a much nicer looking bar than our hotel. We had about half an hour for a drink before dinner. M. ordered a rosé Pinot noir, while I asked for a kolsch beer. The waitress proved to be an Ursula, returning once to ask me what beer it was again, then twice to ask M. whether it was the Pinot noir or the Pinot noir rosé she wanted. And then she brought the non-rosé version and had to change it! And in the middle of all this she took a phone call and chatted for five minutes while we waited without drinks. So it was actually after 19:00, our arranged dinner meeting time, by the time we finished our drinks and left.
The bar was interesting though. They had a VR gaming room right next to it, and you could see people wearing headgear and hand controls doing stuff, and watch their game on a display screen. A guy was fighting virtual ninjas or something, throwing shuriken and shooting arrows at them. Upstairs they had more games, we could see a pool table and a giant vertical pinboard which people fed soccer balls into. Ursula said there was also foosball and other things up there and we should check it out before it closed at 21:00.
 Vegetarian mezze plate at Oasis Grill |
As it turned out, it was okay that we were late to Oasis for dinner, because Mary and Casey didn’t arrive until almost 19:30 because of train delays. We decided to just stay at Oasis because it looked good and there were plenty of tables. M. and I both had a vegetarian mezze plate, which had pita bread, falafels, hummus, tabbouli, tzatziki, baba ganoush, and green salad on it. It was good, and nice to have something salady and healthy. We chatted over dinner until about 21:00, when they decided to head home.
We said goodbye until next time, then walked back to our hotel, stopping in at Walgreens to buy another carton of milk for breakfasts, and also a pack of Pringles for M. to eat on the flight home. I checked how the Chunky Monkey ice cream was going in the hotel room’s bar fridge freezer section. It was half molten, so I decided to I had to finish it off before bed.