Fine-tuning large-scale text-to-video diffusion models to add new generative controls, such as those over physical camera parameters (e.g., shutter speed or aperture), typically requires vast, high-fidelity datasets that are difficult to acquire. In this work, we propose a data-efficient fine-tuning strategy that learns these controls from sparse, low-quality synthetic data. W e show that not only does fine-tuning on such simple data enable the desired controls, it actually yields superior results to models fine-tuned on pho-torealistic "real" data. Beyond demonstrating these results, we provide a framework that justifies this phenomenon both intuitively and quantitatively.

SAG-AFTRA, the union representing Hollywood performers, has reportedly responded to studios' "last, best and final" offer to end the strike, rejecting clauses letting studios re-use AI-created likenesses of high-demand and deceased performers without consent from their estate or families. "They can't have that loophole to exploit performers," a union-side source told The Hollywood Reporter. "We could not allow that language to stand." Reportedly, the Alliance of Motion Picture and Television Producers (AMPTP) would "secure AI scans" for Schedule F performers -- union members earning more than $32,000 per TV episode or $60,000 per film. Studios would pay once to scan the likenesses of these performers without paying for their use or re-use -- essentially giving them eternal rights to their face after paying once upfront.

With the need to feed a growing world population, the efficiency of crop production is of paramount importance. To support breeding and field management, various characteristics of the plant phenotype need to be measured -- a time-consuming process when performed manually. We present a robotic platform equipped with multiple laser and camera sensors for high-throughput, high-resolution in-field plant scanning. We create digital twins of the plants through 3D reconstruction. This allows the estimation of phenotypic traits such as leaf area, leaf angle, and plant height. We validate our system on a real field, where we reconstruct accurate point clouds and meshes of sugar beet, soybean, and maize.

From swooping landscapes to the most dramatic selfie vids ever, a drone can capture it all. These tips will help make your footage extra fly. The first setting to adjust is frame rate: 30 frames per second looks like reality TV, while 24 fps looks like a Hollywood feature--set it to 24. You typically want to shoot in 4K, which will capture the tiniest details and give you flexibility to crop the frame while editing. For action shots, opt for 1080p at 60 or even 120 fps, so you can slow down the footage later.

Let's face it: The AI hype train isn't going away and soon all our devices will be run by artificial intelligence. While Apple's answer to the AI takeover is to just call its new A11 processor "Bionic", Huawei has taken a more concrete approach. The company embedded a neural processing unit (NPU) on its Kirin 970 chip, which it claims can run AI tasks faster and with less power than others. The newly launched Mate 10 Pro is the first phone to use the Kirin 970, and it's meant to demonstrate the wonders of deeply embedded AI. So far though, it's a capable, well-designed phone that has yet to fully explore what a dedicated NPU can do. When Huawei asked a group of reviewers what we wanted from AI, I didn't have a real answer, though my peers pointed out things like natural linguistics and battery management.

To the throngs of those who've put down their complicated digital SLR cameras in favor of easy-to-use smartphones, Ryan Stout has a message: Not so fast. Stout is the founder, CEO and lead developer of Arsenal, a six-person, Montana-based startup that's using computer vision to intelligently automate the abundant capabilities of DSLR cameras. Stout admits he himself had become a smartphone-dependent photographer. But several years ago, he decided to pull his camera out and start taking night photos. From setting shutter speeds, apertures and ISOs to choosing just the right filter, he was quickly reminded that photography is an intensely technical undertaking.

The number of settings on a high-end DSLR or mirrorless camera can be truly overwhelming, even for experienced photographers. That's why Montana-based software developer Ryan Stout created the Arsenal. It's an electronic module that gets plugged into your camera, where it uses artificial intelligence to create the "perfect" shot. Once it's mounted on the hot shoe, plugged in, and wirelessly fired up via an iOS/Android app, the Arsenal assesses the camera's current shot. It does this by comparing it to a database of thousands of other photos in which the camera settings are known, looking for shots that are most similar to the present one.

Relonch wanted me all to fall in love with photography again this CES. But its camera is so radically different from everything I've used before, I struggled to put my faith in its promise. The company is based in Palo Alto, California, and its pitch is simple, if very Silicon Valley: A camera as a service. You hand in your old shooter (yes, really) and in return you get the 291, a unique leather-bound DLSR-shaped camera. It has an APS-C sensor, a fixed, 45mm-equivalent lens, an electronic viewfinder, a shutter key and, importantly, a 4G radio inside.